development of an internal camera-based volume
TRANSCRIPT
![Page 1: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/1.jpg)
University of Arkansas, Fayetteville University of Arkansas, Fayetteville
ScholarWorks@UARK ScholarWorks@UARK
Graduate Theses and Dissertations
12-2017
Development of an Internal Camera-Based Volume Determination Development of an Internal Camera-Based Volume Determination
System for Triaxial Testing System for Triaxial Testing
Sean Elliott Salazar University of Arkansas, Fayetteville
Follow this and additional works at: https://scholarworks.uark.edu/etd
Part of the Civil Engineering Commons, and the Geotechnical Engineering Commons
Citation Citation Salazar, S. E. (2017). Development of an Internal Camera-Based Volume Determination System for Triaxial Testing. Graduate Theses and Dissertations Retrieved from https://scholarworks.uark.edu/etd/2563
This Thesis is brought to you for free and open access by ScholarWorks@UARK. It has been accepted for inclusion in Graduate Theses and Dissertations by an authorized administrator of ScholarWorks@UARK. For more information, please contact [email protected].
![Page 2: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/2.jpg)
Development of an Internal Camera-Based Volume Determination System for Triaxial Testing
A thesis submitted in partial fulfillment
of the requirements for the degree of
Master of Science in Civil Engineering
by
Sean Elliott Salazar
University of Arkansas
Bachelor of Science in Civil Engineering, 2013
December 2017
University of Arkansas
This thesis is approved for recommendation to the Graduate Council.
Dr. Richard Coffman
Thesis Director
Dr. Michelle Bernhardt-Barry Dr. Thomas Oommen
Committee Member Committee Member
![Page 3: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/3.jpg)
ABSTRACT
Accurate measurement of axial, radial, and volumetric strain parameters are critical to
the understanding of phase relationships and the constitutive behavior for saturated and
unsaturated soils. The use of photographic monitoring techniques for laboratory-based
measurement of these parameters have become common. A novel technique that utilized camera
instrumentation located within the triaxial testing cell was developed and validated. By placing
the instrumentation inside of the cell, instead of the instrumentation being located outside of the
cell, the technique eliminated cumbersome corrections required to account for optical distortions
due to 1) the refraction of light at the confining fluid-cell wall-atmosphere interfaces, 2) the
curvature of the cylindrical cell wall, and 3) the deformation of the cell wall induced by changes
in cell pressure. Digital images of various soil and analog (brass, acrylic) specimens were
captured within the triaxial apparatus during testing. The images were processed using the
principles of close-range photogrammetry to construct three-dimensional models of the
specimens. The models were analysed to determine surface deformation and total volume of the
specimens. Additionally, the models obtained from triaxial tests performed on the soil samples
were compared to quantify deformation and volume of the sample as a function of axial strain.
Sensitivity studies and evaluation of measurement accuracy for the internal, close-range
photogrammetry approach are documented herein. Specimen volume, as obtained using the
approach, was compared with volume obtained from four other techniques, including: DSLR
camera photogrammetry, 3D structured light scanning, manual measurements (caliper and pi
tape), and water displacement. A relative error of 0.13 percent was assessed for the internal
photogrammetry technique. The viability of determining total and local strains, volumetric
changes, and total volume at any stage of triaxial testing was demonstrated through undrained
![Page 4: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/4.jpg)
triaxial compression and extension tests. Results from all tests are presented herein. The use of
the internal, close-range photogrammetry technique is recommended.
![Page 5: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/5.jpg)
ACKNOWLEDGEMENTS
A portion of this material is based upon work supported by the National Science
Foundation Graduate Research Fellowship Program under Grant No. DGE-1450079, as awarded
to Sean Salazar. Any opinions, findings, and conclusions or recommendations expressed in this
material are those of the author and do not necessarily reflect the views of the National Science
Foundation. The author would also like to acknowledge 1) support from the University of
Arkansas Provost Collaborative Research Grant, awarded to Richard Coffman (PI), Michelle
Bernhardt (Co-PI) and Adam Barnes (Co-PI), and 2) undergraduate student support from the
United States Department of Transportation (USDOT) Office of the Assistant Secretary for
Research and Technology (OST-R) under Research and Innovation Technology Administration
(RITA) Cooperative Agreement Award No. OASRTRS-14-H-UARK, awarded to Richard
Coffman (PI) and Thomas Oommen (Co-PI).
![Page 6: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/6.jpg)
DEDICATION
I would like to dedicate this thesis to my mother, Betty, to my father, John, to my
grandfather, (Grandpop) Joe, to my brother, Eric, to my sister, Rachel, and to my wife,
Alexandra. Without their love, strength, and unending support, I could not have completed this
work. I would also like to acknowledge the guidance, encouragement, and leadership, of my
advisor, Dr. Richard Coffman. Finally, I would like to thank those colleagues and friends that
helped me along the way, in no particular order, Adam Barnes, Johnathan Blanchard, Cyrus
Garner, Mark Kuss, Nabeel Mahmood, Leah Miramontes, Guilherme Pontes, Morgan Race,
Andrew Schriver, and Brendan Yarborough.
![Page 7: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/7.jpg)
TABLE OF CONTENTS
CHAPTER 1: INTRODUCTION .................................................................................................1
1.1. Chapter Overview .................................................................................................................1
1.2. Description of the Work ........................................................................................................1
1.3. Motivation .............................................................................................................................3
1.3.1. Limitations of Current Techniques .................................................................................3
1.3.2. Contribution to Geotechnical Engineering .....................................................................4
1.4. Document Overview .............................................................................................................5
1.5. References .............................................................................................................................7
CHAPTER 2: BACKROUND ......................................................................................................9
2.1. Chapter Overview .................................................................................................................9
2.2. Parameters of Interest for Triaxial Testing of Soils ..............................................................9
2.3. Non-Photograph-Based Soil Specimen Monitoring Methods .............................................11
2.4. Photograph-Based Soil Specimen Monitoring Methods .....................................................14
2.4.1. Digital Image Analysis Techniques ..............................................................................14
2.4.2. Photogrammetric Techniques .......................................................................................17
2.5. References ...........................................................................................................................24
CHAPTER 3: CONSIDERATION OF INTERNAL BOARD CAMERA OPTICS FOR
TRIAXIAL TESTING APPLICATIONS..................................................................................29
3.1. Chapter Overview ...............................................................................................................29
3.2. Limitations of the Described Study .....................................................................................29
3.3. Consideration of Internal Board Camera Optics for Triaxial Testing Applications ...........30
3.4. Abstract ...............................................................................................................................30
3.5. Introduction .........................................................................................................................31
3.1. Background .........................................................................................................................32
3.6.1. Lens Optics ...................................................................................................................32
3.6.2. Pinhole Aperture ...........................................................................................................33
3.6.3. Pressure Resistant Cameras ..........................................................................................35
3.7. Challenges Encountered ......................................................................................................36
3.7.1. Space Requirements .....................................................................................................37
3.7.2. Confining Fluid ............................................................................................................39
3.7.3. Focal Length .................................................................................................................40
3.7.4. Cell Pressure .................................................................................................................41
3.7.5. Specimen Coverage ......................................................................................................41
3.8. Pinhole Solution ..................................................................................................................43
3.8.1. Video Signal Acquisition .............................................................................................46
3.9. Images Collected .................................................................................................................46
3.10. Conclusions .......................................................................................................................50
3.10.1. Advantages and Limitations of a BCPA ....................................................................51
3.11. References .........................................................................................................................52
![Page 8: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/8.jpg)
CHAPTER 4: DEVELOPMENT OF AN INTERNAL CAMERA-BASED VOLUME
DETERMINATION SYSTEM FOR TRIAXIAL TESTING ..................................................55
4.1. Chapter Overview ...............................................................................................................55
4.2. Limitations of the Described Study .....................................................................................55
4.3. Development of an Internal Camera-Based Volume Determination System for Triaxial
Testing .................................................................................................................................56
4.4. Abstract ...............................................................................................................................56
4.5. Introduction .........................................................................................................................56
4.6. Background .........................................................................................................................57
4.7. Internal Camera-Monitoring System ...................................................................................60
4.7.1. Mechanical Design .......................................................................................................61
4.7.2. Electrical Design ..........................................................................................................62
4.8. Photogrammetric Methods and Results ...............................................................................64
4.9. Conclusions .........................................................................................................................69
4.10. References .........................................................................................................................70
CHAPTER 5: DISCUSSION OF "A PHOTOGRAMMETRY-BASED METHOD TO
MEASURE TOTAL AND LOCAL VOLUME CHANGES OF UNSATURATED SOILS
DURING TRIAXIAL TESTING" BY ZHANG ET AL., 2015 ...............................................73
5.1. Chapter Overview ...............................................................................................................73
5.2. Discussion of "A Photogrammetry-based Method to Measure Total and Local Volume
Changes of Unsaturated Soils During Triaxial Testing" by Zhang et al., 2015 ..................73
5.3. Discussion ...........................................................................................................................74
5.3.1. Improper Triaxial Testing Techniques .........................................................................75
5.3.2. Utilization of Cell Wall Deformation Combined with Least-Square Optimization .....76
5.3.3. Testing Procedures and Photogrammetric Methods .....................................................78
5.4. References ...........................................................................................................................82
CHAPTER 6: CLOSURE TO "DISCUSSION OF 'DEVELOPMENT OF AN INTERNAL
CAMERA-BASED VOLUME DETERMINATION SYSTEM FOR TRIAXIAL TESTING'
BY S. E. SALAZAR, A. BARNES, AND R. A. COFFMAN" BY MEHDIZADEH ET AL.,
2016................................................................................................................................................83
6.1. Chapter Overview ...............................................................................................................83
6.2. Closure to "Discussion of 'Development of an Internal Camera-Based Volume
Determination System for Triaxial Testing' by S. E. Salazar, A. Barnes, and R. A.
Coffman" by Mehdizadeh et al., 2016.................................................................................83
6.3. Abstract ...............................................................................................................................83
6.4. Introduction .........................................................................................................................84
6.5. References ...........................................................................................................................94
![Page 9: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/9.jpg)
CHAPTER 7: VERIFICATION OF AN INTERNAL CLOSE-RANGE
PHOTOGRAMMETRY APPROACH FOR VOLUME DETERMINATION DURING
TRIAXIAL TESTING .................................................................................................................96
7.1. Chapter Overview ...............................................................................................................96
7.2. Limitations of the Described Study .....................................................................................97
7.3. Verification of an Internal Close-Range Photogrammetry Approach for Volume
Determination During Triaxial Testing ...............................................................................97
7.4. Abstract ...............................................................................................................................97
7.5. Introduction and Background ..............................................................................................98
7.6. Evaluation of the Internal Photogrammetry Technique ....................................................101
7.6.1. Calibration of Board Cameras ....................................................................................104
7.6.2. Derivation of Camera Stations within the Triaxial Cell .............................................105
7.6.3. Determination of Photograph-Capturing Intervals .....................................................106
7.6.4. Capture of Photographs to Determine Accuracy in Confining Fluid .........................107
7.6.5. Photogrammetric Reconstruction of a Specimen .......................................................108
7.6.6. Determination of a Specimen Volume .......................................................................110
7.6.7. Accuracy of Technique ...............................................................................................110
7.6.7.1. DSLR Camera Photogrammetry .........................................................................111
7.6.7.2. 3D Scanning ........................................................................................................112
7.6.7.3. Manual Measurements ........................................................................................113
7.6.7.4. Water Displacement ............................................................................................114
7.6.8. Limitations and Sources of Error ...............................................................................114
7.6.8.1. Precision of Repeat Interval Stops ......................................................................115
7.6.8.2. Model Refinement ..............................................................................................116
7.6.8.3. External Geometry Measurements ......................................................................116
7.6.8.4. Determination of Specimen Ends .......................................................................116
7.7. Implementation of the Internal Photogrammetry Technique for Soil Specimens .............117
7.8. Results and Discussion ......................................................................................................121
7.8.1. Volume Comparisons .................................................................................................121
7.8.2. Photograph Interval ....................................................................................................123
7.8.3. Testing of Internal Photogrammetry System on Soil Specimens ...............................124
7.9. Conclusions .......................................................................................................................129
7.9.1. Potential Applications and Future Improvements ......................................................130
7.10. Acknowledgements .........................................................................................................131
7.11. References .......................................................................................................................132
CHAPTER 8: CONCLUSIONS ...............................................................................................135
8.1. Chapter Overview .............................................................................................................136
8.2. Highlights ..........................................................................................................................136
8.3. Limitations ........................................................................................................................137
8.4. Recommendations .............................................................................................................137
CHAPTER 9: WORKS CITED ................................................................................................139
![Page 10: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/10.jpg)
LIST OF FIGURES
Figure 1.1. Exploded, transparent view of the major components of the internal camera-based
photogrammetry system (note: shown with piezoelectric end caps)…………………...…2
Figure 1.2. Overview of the evaluation process for the internal cell photogrammetry technique
presented in this document (modified from Salazar et al. 2017b)…………………….......3
Figure 2.1. Typical measurements and calculations required to determine the phase diagram of
the soil specimen during a reduced triaxial extension test (modified from Salazar et al.
2017b)…………………………………………………………………............................11
Figure 2.2. Schematic of a double-cell volume measuring system for triaxial testing of
unsaturated soils (from Ng et al. 2002)…………….………………………………….....12
Figure 2.3. Plan view schematic of laser scanning device for triaxial testing (from Messerklinger
and Springman 2007)………………………………………………………………….....14
Figure 2.4. Triaxial testing apparatus with (a) a single, fixed digital camera located outside of the
cell (from Gachet et al. 2007), and (b) three, fixed digital cameras located outside of the
cell (from Bhandari et al. 2012)……………………………………………………….....17
Figure 2.5. Optical magnification of soil specimen due to presence of confining fluid…………18
Figure 2.6. (a) Schematic of photogrammetric principles involved in the Zhang et al. (2015) and
Li et al. (2016) methodology, including (b) optical ray tracing, and (c) least-square
estimation (from Li et al. 2016)……………………………………………………….....20
Figure 2.7. Representation of specimen deformations obtained during a triaxial test (modified
from Li et al. 2016)………………………………………………………………………20
Figure 2.8. Schematic of board camera device with pinhole aperture developed at the University
of Arkansas (from Salazar and Coffman 2015)……………………………………….....22
Figure 2.9. Schematic of internal camera-based photogrammetry system for triaxial testing
developed at the University of Arkansas (modified from Salazar et al. 2015)…………..22
Figure 2.10. Representation of specimen deformation obtained from close-range
photogrammetry during an undrained, conventional, triaxial compression test (modified
from Salazar et al. 2017b)………………………………………………………………..23
Figure 3.1. Real image formation illustrated by simplified ray diagram of (a) diffraction of light
through a pinhole aperture, and (b) refraction of light through a lens…………………...35
Figure 3.2. The process that was followed to address the interrelated challenges in the design of
the BCPA………………………………………………………………………………...37
![Page 11: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/11.jpg)
Figure 3.3. (a) Three types of board cameras that were tested, and (b) five different types of
lenses used with the cameras…………………………………………………………….39
Figure 3.4. Schematic of guided camera track system mounted on triaxial apparatus base……..42
Figure 3.5. Schematic of (a) front view, (b) exploded side view, and (c) exploded orthogonal
view of the BCPA………………………………………………………………………..45
Figure 3.6. Photograph of one of the BCPA utilized inside of the triaxial cell……………….....45
Figure 3.7. Still frames captured using a 8.38mm (0.33 in.) format CCD board camera with
(1) 3.7mm button lens (55º FOV, M12×0.5 thread) in (a) air, and within (b) PSF-5cSt
silicone oil, and with (2) 75 μm diameter pinhole aperture (attached to a M12×0.5 barrel
with 6.5mm diameter opening) in (c) air, and within (d) PSF-5cSt silicone oil…………47
Figure 4.1. (a) Exploded view, and (b) elevation view of the internal components of the
combined BCPA monitoring system………………………………………………….....61
Figure 4.2. Photograph of the power switchboard for the BCPA devices (power supply and
timing sequence)…………………………………………………………………………63
Figure 4.3. Wiring diagram of the photogrammetric instrumentation…………………………...64
Figure 4.4. (a) PhotoModeler camera calibration grid, (b) DSLR-acquired survey images
(control point identification), (c) BCPA-acquired calibration images (camera location
and orientation identification), and (d) BCPA-acquired images (point cloud
identification)………………………………………………………………………….....66
Figure 4.5. Vertically and horizontally overlapping photographs captured with two adjacent
BCPA on a single tower at 15º rotation intervals………………………………………..66
Figure 4.6. Photogrammetric reconstruction of BCPA locations within the triaxial cell for
(a) 45º, (b) 30º, (c) 15º, and (d) 5º rotation intervals………………………………...…..68
Figure 6.1. Comparison of deviatoric stress as a function of mean stress for a reduced triaxial
extension test with pauses and without pauses for capturing photographs………………91
Figure 6.2. Comparison of deviatoric stress as a function of axial strain for a reduced triaxial
extension test with pauses and without pauses for capturing photographs………………91
Figure 6.3. Comparison of excess pore water pressure development as a function of axial strain
for a reduced triaxial extension test with pauses and without pauses for capturing
photographs…………………………………………………………………………........92
Figure 7.1. The process used to evaluate the internal cell photogrammetry technique and to
obtain test parameters………………………………………………………………..…102
![Page 12: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/12.jpg)
Figure 7.2. The process used to determine the volume of a specimen using internal cell
photogrammetry and to evaluate the sensitivity of photograph interval on the volume
of the specimen………...……………………………………………………………….103
Figure 7.3. (a) Photograph of, and (b) three-dimensional, watertight model of small-acrylic
analog specimen with spherical adhesive targets (removed during processing), as
obtained during 3D scanning of specimen………………………………………...........113
Figure 7.4. Qualitative factors affecting accuracy in photogrammetry (modified from Eos
Systems, Inc. 2015)……………………………………………………………………..115
Figure 7.5. Typical measurements and calculations required to determine the phase diagram of
the soil specimen during for conventional triaxial compression test………...................118
Figure 7.6. Photograph of the kaolinite specimen within the photogrammetrically instrumented
triaxial cell during the shearing stage of the extension test…………………………….120
Figure 7.7. Sensitivity of derived camera location difference to interval between
photographs……………………………………………………………………………..124
Figure 7.8. Visualization of photogrammetry-obtained, three-dimensional models of kaolinite
specimen during K0-consolidation phase of triaxial test (warm colors indicate positive
deformation and cool colors indicate negative deformation)…………………………..127
Figure 7.9. Visualization of photogrammetry-obtained, three-dimensional models of kaolinite
test specimen during (a) conventional triaxial compression, and (b) reduced triaxial
extension tests up to 15 percent axial strain during shearing (warm colors indicate
positive deformation and cool colors indicate negative deformation)……………….....128
![Page 13: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/13.jpg)
LIST OF TABLES
Table 3.1. Photogrammetric properties of the BCPA……………………………………………49
Table 7.1. Comparison of intrinsic camera parameters from calibration for all ten board cameras
utilized for internal photogrammetry…………………………………………………...105
Table 7.2. Comparison of small-acrylic analog specimen volumes as obtained using five different
techniques………………………………………………………………………………122
Table 7.3. Comparison of large-acrylic analog specimen volumes as determined during internal
photograph interval sensitivity test……………………………………………………..123
Table 7.4. Volumes of kaolinite soil specimens as determined throughout the triaxial
compression test and corresponding summary statistics……………………………….126
Table 7.5. Volumes of kaolinite soil specimen as determined throughout the triaxial extension
test and corresponding summary statistics……………………………………………...126
Table 7.6. Volumes of kaolinite soil specimen as determined throughout the unconfined
compression test and corresponding summary statistics……………………………….126
![Page 14: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/14.jpg)
LIST OF SYMBOLS AND ACRONYMS
Symbol Definition
3D Three-dimensional
Ac Corrected area
Af Area of failure plane
BCPA Board Camera with Pinhole Aperture
CCD Charge-Coupled Device
cm Centimeter
CMOS Complementary Metal Oxide Semiconductor
CTC Conventional Triaxial Compression
DIA Digital Image Analysis
DIC Digital Image Correlation
DSLR Digital Single Lens Reflex
d Diameter
d0 Initial diameter
f Focal length
h0 Initial height
K0 Coefficient of lateral earth pressure at-rest
K1, K2, K3 Dimensionless coefficients of radial distortion
kPa Kilopascal
LoHR Lines of Horizontal Resolution
m Mass
mm Millimeter
MVS Multi-View Stereo
n1 Refractive index of lens material
n2 Refractive index of surrounding medium
OCR Overconsolidation Ratio
p' Mean stress
P1, P2 Dimensionless coefficients of decentering distortion
PIV Particle Image Velocimetry
psi Pounds per square inch
q Deviatoric stress
r Radius of the pinhole opening (aperture)
r1 Radius of curvature of front surface of lens
r2 Radius of curvature of back surface of lens
RAD Ringed Automatically Detected
RTE Reduced Triaxial Extension
SfM Structure from Motion
UC Unconfined Compression
V Volume
V0 Initial volume
VT Total volume
w Water content
wc Water content
Δd Change in diameter
![Page 15: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/15.jpg)
Δh Change in height
Δu Excess pore pressure
ΔV Change in volume
εa Axial strain
εr Radial strain
εv Volumetric strain
λ Design wavelength
μm Micrometer
σ1 Major principal stress
σ3 Confining stress
º Degree
% Percent
![Page 16: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/16.jpg)
LIST OF PUBLISHED OR SUBMITTED PAPERS
Chapter 3: Consideration of Internal Board Camera Optics for Triaxial Testing Applications.
Published as: Salazar, S.E., Coffman, R.A. (2015). “Consideration of Internal Board
Camera Optics for Triaxial Testing Applications.” Geotechnical Testing Journal,
Vol. 38, No. 1, pp. 40-49. doi:10.1520/GTJ20140163.
Chapter 4: Development of an Internal Camera-Based Volume Determination System for
Triaxial Testing. Published as: Salazar, S.E., Barnes, A., Coffman, R.A. (2015).
“Development of an Internal Camera-Based Volume Determination System for Triaxial
Testing.” Geotechnical Testing Journal, Vol. 38, No. 4, pp. 549-555.
doi:10.1520/GTJ20140249.
Chapter 5: Discussion of “A Photogrammetry-based Method to Measure Total and Local
Volume Changes of Unsaturated Soils During Triaxial Testing” by Zhang et al., 2015.
Published as: Salazar, S.E., Coffman, R.A. (2015). “Discussion of ‘A Photogrammetry-
based Method to Measure Total and Local Volume Changes of Unsaturated Soils During
Triaxial Testing’ by Zhang et al.” Acta Geotechnica, Vol. 10, No. 5, pp. 693-696.
doi:10.1007/s11440-015-0380-1.
Chapter 6: Closure to “Discussion of ‘Development of an Internal Camera-Based Volume
Determination System for Triaxial Testing’ by S. E. Salazar, A. Barnes, and R. A.
Coffman” by Mehdizadeh et al., 2016. Published as: Salazar, S.E., Barnes, A., Coffman,
R.A. (2017). “Closure to “Discussion of ‘Development of an Internal Camera-Based
Volume Determination System for Triaxial Testing’ S.E. Salazar, A. Barnes, and R.A.
Coffman” by A. Mehdizadeh, M.M. Disfani, R. Evans, A. Arulrajah and D.E.L. Ong.”
Geotechnical Testing Journal. Vol. 40, No. 1, pp. 47-51. doi:10.1520/GTJ20160154.
Chapter 7: Verification of an Internal Close-Range Photogrammetry Approach for Volume
Determination During Triaxial Testing. Submitted as: Salazar, S.E., Miramontes, L.D.,
Barnes, A.R., Bernhardt, M.L., Coffman, R.A. (2017). “Verification of an Internal Close-
Range Photogrammetry Approach for Volume Determination During Triaxial Testing.”
Geotechnical Testing Journal. Submitted for Review. Manuscript Number: GTJ-2017-
0125.R2.
![Page 17: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/17.jpg)
1
CHAPTER 1: INTRODUCTION
1.1. Chapter Overview
The development of an internal camera-based volume determination system for triaxial
testing is described in this document. The system was used in conjunction with a close-range
photogrammetry technique to 1) capture digital images, 2) photogrammetrically construct three-
dimensional models, and 3) calculate the volume and deformation for soil specimens during all
stages of triaxial compression and triaxial extension tests. This chapter is subdivided into four
sections. A brief overview of the work that is described in this document is contained in Section
1.2. The motivation for conducting this research is described in Section 1.3. An overview of the
entire document is presented in Section 1.4.
1.2. Description of the Work
The development of the internal camera-based instrumentation for triaxial cell
photogrammetry and the associated data collection and processing techniques are described in
this document. The development is also briefly described in this section. A set of ten small board
cameras was designed and incorporated into two, diametrically opposed, camera towers. Each
tower, with five camera devices, was mounted to a rotational track within the cell; the towers
were free to rotate about the soil specimen during triaxial testing. At any desired stage during the
triaxial test (e.g. at a given axial strain level during shearing), the test was paused and the camera
towers were rotated incrementally about the specimen while capturing images of the specimen.
To facilitate measurements during the triaxial test, the entire system was designed to be
incorporated into the triaxial cell to be in direct contact with the confining fluid. A rendering of
the camera instrumentation, as mounted inside of the triaxial cell, is presented as Figure 1.1.
![Page 18: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/18.jpg)
2
Figure 1.1. Exploded, transparent view of the major components of the internal camera-
based photogrammetry system (note: shown with piezoelectric transducer end caps).
The images that were collected with the system were processed using a close-range
photogrammetry technique to 1) construct the digital surface of the specimen, and 2) determine
the total volume of the specimen. To demonstrate the viability of the technique for triaxial tests,
one conventional triaxial compression test and one reduced triaxial extension test was performed.
During each test, images of the specimen were captured for various levels of axial strain and a
three-dimensional model was created for each of the various axial strain levels. To evaluate the
accuracy of the internal photogrammetry technique, several validation tests were performed.
Specifically, the technique was evaluated using analog specimens (one brass and two acrylic
specimens). The effect of the number of images on the reconstruction of a specimen (ranging
from 40 to 320 images) was examined and the total volume of the specimens that were obtained
Triaxial testing cell base
Cell wall
Individual board camera device
Soil specimen with photogrammetric targets on membrane
Camera tower
Rotating track
Magnet
![Page 19: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/19.jpg)
3
were compared with volumes measured using four other methods. An overview of the evaluation
steps is presented as Figure 1.2.
Figure 1.2. Overview of the evaluation process for the internal cell photogrammetry
technique presented in this document (modified from Salazar et al. 2017b).
1.3. Motivation
The motivation for the research conducted for this work is presented in this section. The
limitations of the techniques that are currently employed in the laboratory to monitor triaxial
specimens during testing are discussed in Section 1.3.1. The contribution of the work to the field
of geotechnical engineering is discussed in Section 1.3.2.
1.3.1. Limitations of Current Techniques
Various methods have been employed by researchers to monitor soil specimens during
triaxial tests. These monitoring efforts have enabled one or more of the following: 1)
measurement of axial and/or radial dimensions and deformations during testing, 2) measurement
Volu
mes C
alc
ula
ted
3-D Scan
DSLR Camera Photogrammetry
Manual Measurements
Water Displacement
Comparison of Volumes
CTC and RTE Testing of Soil Specimens
Capture photos of specimen with camera towers at 20° rotation interval
for any given stage of testing
DSLR Target Identification
Internal Camera Location/Orientation
Create 3-D volumes of specimen for any given stage of testing
Calculate specimen parameters Visualize strains
Internal Cell Photogrammetry Evaluation
Volume Determination
Method Comparison
Internal Cell Photogrammetry
![Page 20: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/20.jpg)
4
of local and/or total volume, 3) calculation of axial, radial, and/or volumetric strains, and 4)
characterization of shear banding behavior. Techniques that have been described in the literature
include double-wall cell systems (e.g. Bishop and Donald 1961), differential pressure transducers
(e.g. Ng et al. 2002), measurements of air and water volume changes (e.g. Leong et al. 2004),
displacement and strain sensors (e.g. Scholey et al. 1995), non-contact proximity sensors (e.g.
Clayton et al. 1989), laser scanners (e.g. Messerklinger and Springman 2007), digital image
analysis (e.g. Bhandari et al. 2012), x-ray computed tomography (e.g. Viggiani et al. 2004), and
photogrammetry (e.g. Zhang et al. 2015). Broadly speaking, these techniques can be divided into
photograph-based and non-photograph-based categories.
In recent years, photograph-based methods have achieved prominence due to their
practicality, cost-effectiveness, and versatility. Furthermore, many of the non-photograph-based
techniques suffered from poor data resolution, relied heavily on geometric assumptions, and
often required installation of complex and expensive instrumentation. However, even the
photograph-based techniques have been limited by poor resolution, and the need to perform
computationally intensive corrections to overcome distortions caused by the confining fluid and
the cell wall surrounding the soil specimen. The limitations of the photograph-based and non-
photograph-based approaches were discussed in detail in Salazar and Coffman (2015a, 2015b)
and Salazar et al. (2015, 2017a, 2017b). Based on a review of the existing literature, there was a
need for a better photogrammetry technique that relied upon cameras internal to the cell.
1.3.2. Contribution to Geotechnical Engineering
Testing of the novel Salazar and Coffman (2016) device, in conjunction with the close-
range photogrammetry technique detailed in Salazar et al. (2017b), was shown to be a viable
alternative to other photograph-based techniques. Moreover, the technique allowed for direct
![Page 21: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/21.jpg)
5
observation and coverage of the entire radial surface of a specimen. This coverage resulted in
highly detailed construction of the three-dimensional surface of the specimen. Moreover, the
measurements were independent of any assumptions of initial dimensions or deformation
behavior. The technique simplified the computations required by other photogrammetric
methods, such as the method presented by Zhang et al. (2015). The potential for obtaining more
information about a given soil specimen, such as axial, radial, and volumetric strains during a
triaxial test is demonstrated by the technique that is presented in this document. This information
can be used together with the other soil parameters to improve the development of constitutive
models of soil. With continued improvements to the device and to the processing workflow, the
potential for less time-consuming data collection and data reduction is envisioned.
1.4. Document Overview
This document is comprised of nine chapters. In this chapter (Chapter 1), an overview of
the work contained within the document and the contribution of the work to the field of
geotechnical engineering were provided. The background for the work is presented in the form
of a literature review in Chapter 2. Five subsequent archival journal publications, on the subject
of this work, are presented in Chapters 3 through 7, in the order in which the manuscripts were
conceived and published. The process of developing suitable cameras for the internal
photogrammetry system is described in Chapter 3. The development of the mechanical,
electrical, and photogrammetric components of the system is described in Chapter 4. A
discussion of the paper by Zhang et al. (2015), on the topic of photogrammetry for triaxial
testing, is included as Chapter 5. A closure to a discussion paper written by Mehdizadeh et al.
(2016) on the Salazar et al. (2015) publication is included as Chapter 6. The validation of the
internal, close-range photogrammetry approach for triaxial testing is described in Chapter 7.
![Page 22: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/22.jpg)
6
Conclusions drawn from the work presented in Chapter 3 through Chapter 7 are discussed in
Chapter 8. Finally, a comprehensive list of works cited in this document is included as Chapter 9.
![Page 23: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/23.jpg)
7
1.5. References
Bhandari, A. R., Powrie, W., and Harkness, R. M., 2012, “A Digital Image-Based Deformation
Measurement System for Triaxial Tests,” Geotech. Test. J., Vol. 35, No. 2, pp. 1-18.
Bishop, A. W. and Donald, I. B., 1961, “The Experimental Study of Partly Saturated Soil in
Triaxial Apparatus,” Proceedings of the Fifth International Conference on Soil
Mechanics and Foundation Engineering, Vols. 1/3, Paris, July 17–22, Institution of Civil
Engineers, London, pp. 13–21.
Clayton, C. R. I., Khatrush, S. A., Adriano, V. D. B., and Siddique, A., 1989, “The Use of Hall
Effect Semiconductors in Geotechnical Instrumentation,” Geotech. Test. J., Vol. 12,
No. 1, pp. 69-76.
Kikkawa, N., Nakata, Y., Hyodo, M., Murata, H., and Nishio, S., 2006, "Three-Dimensional
Measurement of Local Strain Using Digital Stereo Photogrammetry in the Triaxial Test,"
Geomechanics and Geotechnics of Particulate Media, M. Hyodo, H. Murata, and Y.
Nakata, Eds., Taylor and Francis, London, pp. 61-67.
Leong, E.C., Agus, S.S., and Rahardjo, H., 2004, “Volume Change Measurement of Soil
Specimen in Triaxial Test,” Geotech. Test. J., Vol. 27, No. 1, pp. 47-56.
Mehdizadeh, A., Disfani, M. M., Evans, R., Arulrajah, A., and Ong, D. E. L., 2016, “Discussion
of ‘Development of an Internal Camera-Based Volume Determination System for
Triaxial Testing’ by S. E. Salazar, A. Barnes and R. A. Coffman
(doi:10.1520/GTJ20140249).” Geotech. Test J., Vol. 39, No. 1, pp. 165-168.
doi:10.1520/GTJ20150153.
Messerklinger, S. and Springman, S. M., 2007, “Local Radial Displacement Measurements of
Soil Specimens in a Triaxial Test Apparatus Using Laser Transducers,” Geotech. Test. J.,
Vol. 30, No. 6, pp. 1-12.
Ng, C. W. W., Zhan L. T., and Cui Y. J., 2002, “A New Simple System for Measuring Volume
Changes in Unsaturated Soils,” Can. Geotech. J., Vol. 39, No. 3, pp. 757-764.
Salazar, S. E., Barnes, A., and Coffman, R. A., 2015, “Development of an Internal Camera
Based Volume Determination System for Triaxial Testing,” Geotech. Test. J., Vol. 38,
No. 4, pp. 549-555. doi:10.1520/GTJ20140249.
Salazar, S. E., Barnes, A., and Coffman, R. A., 2017a, “Closure to “Discussion of ‘Development
of an Internal Camera-Based Volume Determination System for Triaxial Testing’ S. E.
Salazar, A. Barnes and R. A. Coffman” by A. Mehdizadeh, M. M. Disfani, R. Evans, A.
Arulrajah and D. E. L. Ong,” Geotech. Test. J., Vol. 40, No. 1, pp. 47-51.
doi:10.1520/GTJ20160154.
![Page 24: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/24.jpg)
8
Salazar, S. E. and Coffman, R. A., 2015a, “Consideration of Internal Board Camera Optics for
Triaxial Testing Applications,” Geotech. Test. J., Vol. 38, No. 1, pp. 40-49.
doi:10.1520/GTJ20140163.
Salazar, S. E. and Coffman, R. A., 2015b, “Discussion of “A Photogrammetry-Based Method to
Measure Total and Local Volume Changes of Unsaturated Soils During Triaxial Testing”
by Zhang et al. (doi:10.1007/s11440-014-0346-8)” Acta Geotech., Vol. 10, No. 5,
pp. 693-696. doi:10.1007/s11440-015-0380-1.
Salazar, S. E. and Coffman, R. A., 2016, “Pressurized Fluid-Submerged, Internal, Close-Range
Photogrammetry System for Laboratory Testing,” U.S. Patent and Trademark Office.
Patent 15/360,820. Filed November, 2016.
Salazar, S. E., Miramontes, L. D., Barnes, A., Bernhardt, M. L., and Coffman, R. A., 2017b,
“Verification of an Internal Close-Range Photogrammetry Approach for Volume
Determination During Triaxial Testing,” Geotech. Test. J. Submitted for Review.
Manuscript Number: GTJ-2017-0125.R2.
Scholey, G. K., Frost, J. D., Lo Presti, C. F., and Jamiolkowski, M., 1995, “A Review of
Instrumentation for Measuring Small Strains During Triaxial Testing of Soil Specimens,”
Geotech. Test. J., Vol. 18, No. 2, pp. 137–156.
Viggiani, G., Lenoir, N., Bésuelle, P., Di Michiel, M., Marello, S., Desrues, J., and Kretzschmer,
M., 2004, “X-ray Microtomography for Studying Localized Deformation in Fine-Grained
Geomaterials Under Triaxial Compression,” Cr Acad Sci II B-Mec, Vol. 332,
pp. 819-826.
Zhang, X., Li, L., Chen, G., and Lytton, R., 2015, “A Photogrammetry-Based Method to
Measure Total and Local Volume Changes of Unsaturated Soils During Triaxial
Testing,” Acta Geotechnica, Vol. 10, No. 1, pp. 55-82. doi:10.1007/s11440-014-0346-8.
![Page 25: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/25.jpg)
9
CHAPTER 2: BACKGROUND
2.1. Chapter Overview
A review of the relevant literature is contained within this chapter. The parameters of
interest for triaxial testing of saturated and unsaturated soils are presented in Section 2.2. An
overview of non-photograph-based soil specimen monitoring methods, with a focus on triaxial
testing, is provided in Section 2.3. A summary of the literature related to photograph-based soil
specimen monitoring, with a focus on triaxial testing, is presented in Section 2.4. Subsections for
digital image analysis techniques (Section 2.4.1) and photogrammetric techniques (Section 2.4.2)
are contained within Section 2.4.
2.2. Parameters of Interest for Triaxial Testing of Soils
Numerous techniques have been applied to study the volume and strain evolutions for
saturated and unsaturated soil specimens during triaxial tests. The practice of measuring changes
in volume during a test has become routine for many laboratories and is critical for triaxial
testing of unsaturated soils. The parameters of interest for a given triaxial test typically include:
1) total and local volume changes, and 2) axial, radial, and volumetric strains. Although there are
several methods for indirectly obtaining or calculating these parameters, assumptions of elastic
and uniform deformation behaviors to estimate shape are typically associated with this approach.
An example of one set of calculations for obtaining axial, radial, and volumetric strains during an
undrained triaxial test is presented in Equations 2.1, 2.2, and 2.3, respectively. This method relies
on a geometric estimation of specimen shape after deformation and relies on relative changes in
specimen dimensions. Ehrgott (1971) presented five additional variations of the equations used
to calculate the strains, but all of the variations suffer from fundamentally flawed assumptions
regarding the specimen shape.
![Page 26: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/26.jpg)
10
𝜀𝑎 =𝛥ℎ
ℎ0 (modified from Ehrgott 1971) Equation 2.1
𝜀𝑟 =𝛥𝑑
𝑑0 (modified from Ehrgott 1971) Equation 2.2
𝜀𝑣 =𝛥𝑉
𝑉0= 𝜀𝑎 + 𝜀𝑟 − 𝜀𝑟𝜀𝑎 +
𝜀𝑟2
3(𝜀𝑎 − 1) (modified from Ehrgott 1971) Equation 2.3
Where εa is axial strain, Δh and h0 are change in height and initial height of a test specimen,
respectively, εr is radial strain, Δd and d0 are change in diameter and initial diameter of a test
specimen, respectively, εv is volumetric strain, and ΔV and V0 are change in volume and initial
volume of a test specimen, respectively.
To further illustrate an example of the assumptions that are typically made regarding the
specimen deformation behavior during a triaxial test, the five testing stages of a reduced triaxial
extension test are illustrated in Figure 2.1. It is commonly assumed that the specimen deforms as
a perfect, right, circular cylinder (ASTM D4767, 2011). Therefore, this technique is not suitable
for obtaining the correct area of the failure plane and the resulting shear strength calculations for
the specimen are inaccurate. Due to these inaccuracies, researchers have turned to a variety of
techniques to directly measure the parameters of interest during triaxial testing.
![Page 27: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/27.jpg)
11
1. Pre-test: Mass (m) and water content (w), measured; Volume (V) calculated using caliper measurements. 2. Back-pressure saturation: Drain lines filled. Total volume change (ΔV) from pore pump measurements. This volume change includes air 1) purged from lines, and 2) going into suspension. 3. K0 Consolidation: Sample ΔV from pore pump measurements. 4. Shearing: m, w, and V assumed to be equal to post-test m, w, and V (if undrained); calculated from pore pump measurements (if drained). 5. Post-test: m and w, measured. Shear strength determined based on corrected area (Ac).
Figure 2.1. Typical measurements and calculations required to determine the phase
diagram of the soil specimen during a reduced triaxial extension test (modified from
Salazar et al. 2017b).
2.3. Non-Photograph-Based Soil Specimen Monitoring Methods
Historically, changes in confining fluid or pore fluid volume have been directly
correlated with changes in specimen volume. However, volume measurements have been
influenced by temperature- and pressure-induced flexure of the cell wall and drain lines. Bishop
and Donald (1961) modified a standard triaxial testing apparatus to include a second, inner cell
that was filled with mercury to measure total changes in volume of a specimen. Other double-cell
techniques that relied on measuring the changes in volume of the confining fluid (water and/or
air) within the pressurized cell were introduced by Wheeler (1986), Cui and Delage (1996),
Toyota et al. (2001), Aversa and Nicotera (2002), and Ng et al. (2002). A review of these volume
measurement techniques was provided in Leong et al. (2004), which also contained a method for
correcting for the expansion of the confining fluid due to temperature fluctuations. An example
of a double-cell triaxial apparatus is presented as Figure 2.2.
1 2 3 4 5
![Page 28: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/28.jpg)
12
To complement volume measurements, axial deformation measurements have also been
collected during testing. Changes in axial deformation have been used to calculate average
specimen dimensions by adding or subtracting deformation measurements from initial
dimensions, with the initial dimensions having been established prior to testing (typically by
means of caliper and pi tape measurements). Therefore, known or assumed specimen dimensions
were used to calculate axial, radial, or volumetric strains, as described in Section 2.2. However,
conventional axial (Scholey et al. 1995, Cuccovillo and Coop 1997, Ng and Chiu 2001) and
lateral or radial (Khan and Hoag 1979, Bésuelle and Desrues 2001) displacement transducer
measurements have relied on averaging methods and total volumetric changes that did not
accurately account for irregular deformation behavior, such as shear banding, bulging
bifurcation, or necking.
Figure 2.2. Schematic of a double-cell volume measuring system for triaxial testing of
unsaturated soils (from Ng et al. 2002).
![Page 29: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/29.jpg)
13
Other measurement techniques have been employed to measure soil specimen
parameters. Clayton and Khatrush (1986) and Clayton et al. (1989) introduced a non-contact
proximity sensor technique to measure local radial and axial strains during triaxial testing.
Romero et al. (1997), Messerklinger and Springman (2007), and Jain et al. (2015) employed
laser-scanning devices. The aforementioned Romero et al. (1997) device was incorporated into a
suction- and temperature-controlled triaxial apparatus for the testing of unsaturated soils. Two
externally mounted, diametrically opposed lasers were utilized to measure the radial deformation
profile along the length of the specimen at two locations. Similarly, the Messerklinger and
Springman (2007) device was utilized to measure radial displacements for three vertical profiles
around the circumference of the specimen during triaxial testing (Figure 2.3). In both techniques,
the radial displacements between the measured profiles were inferred. Although the Jain et al.
(2015) device was not employed during triaxial tests, the technique utilized a fixed laser pointed
at a soil specimen placed on a rotating turntable. This allowed for continuous measurements of
the entire specimen surface at desired intervals during a shrinkage test and subsequent
calculation of specimen volume. Desrues et al. (1996) and Viggiani et al. (2004) utilized x-ray
computed tomography to study the triaxial soil specimens.
![Page 30: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/30.jpg)
14
Figure 2.3. Plan view schematic of laser scanning device for triaxial testing (from
Messerklinger and Springman 2007).
2.4. Photograph-Based Soil Specimen Monitoring Methods
To overcome the limitations of the previously discussed techniques in Section 2.3, digital
imaging techniques have increasingly been employed to monitor soil specimens. Specifically,
these techniques have been used to 1) calculate the total volume and volumetric strain of soil
specimens and/or to 2) monitor the evolution of shear bands and local strains. The photograph-
based techniques have been shown to be robust alternatives to conventional measurement
techniques for obtaining measurements during triaxial testing. The photograph-based techniques
in the literature fall under one or more of the following classifications: Digital Image Analysis
(DIA), Digital Image Correlation (DIC), Particle Image Velocimetry (PIV), or photogrammetry.
For the purposes of this discussion, DIA, DIC, and PIV techniques have all been grouped under
the digital image analysis category, while photogrammetry technique is treated separately.
2.4.1. Digital Image Analysis Techniques
Throughout the literature, the terms DIA and DIC have sometimes been used
interchangeably and the distinctions of the methods are beyond the scope of this discussion.
![Page 31: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/31.jpg)
15
Therefore, these methods are discussed together and are referred to collectively as DIA
techniques herein. DIA techniques have allowed for more information to be captured and
quantified for soils specimens. For example, Alshibli and Sture (1999) utilized a uniform grid
applied to the membrane of a triaxial soil specimen to measure displacements at the surface of
the specimen during the development of a shear band. In another set of examples, Gudehus and
Nübel (2004) and Rechenmacher and Finno (2004) both studied the evolution of shear bands in
sands during biaxial tests using digital image analysis. In yet another pair of examples, Ören et
al. (2006) and Önal et al. (2008) used digital image correlation to determine the volume of soil
specimens during shrinkage tests.
Another class of digital image analysis is PIV; PIV was developed by Adrian (1991) for
experimental fluid dynamics applications. Although PIV techniques share some common traits
with DIA techniques, PIV differs significantly from DIA. PIV techniques primarily utilize image
texture instead of target markers to track movement in sequential images. PIV has been used to
measure planar surface deformation and to analyze displacement and strain fields for various soil
tests. Examples of PIV techniques that have been employed to monitor soil specimens include
Guler et al. (1999), White et al. (2003), Iskander and Liu (2010), Stanier et al. (2016), and Pinyol
and Alvarado (2017). Because PIV techniques have not been shown to aid in the determination
of triaxial specimen volumes, these techniques are not further discussed.
The use of digital images in conjunction with DIA techniques for monitoring triaxial tests
has been presented in the literature. Examples include Macari et al. (1997), Alshibli and Sture
(1999), Alshibli and Al-Hamdan (2001), Gachet et al. (2007), Sachan and Penumadu (2007),
Rechenmacher and Medina-Cetina (2007), Uchaipichat et al. (2011), Bhandari et al. (2012), and
Hormdee et al. (2014). In each of these examples, digital images of the soil specimens were
![Page 32: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/32.jpg)
16
captured during testing with photographic equipment that was located outside of the triaxial
testing cell. Although the entire length of the specimen (in the axial dimension) was captured
within a single image in each of the aforementioned references, various methods were used to
capture the entire surface area of the specimen in the lateral dimension. For example, Alshibli
and Al-Hamdan (2001) and Bhandari et al. (2012) placed multiple cameras at intervals around
the outside of the cell, whereas Macari et al. (1997) and Gachet et al. (2007) utilized only a
single, fixed camera and therefore did not capture the entire specimen surface. In other instances,
the entire specimen surface was not captured because only the zone of shear banding was of
interest (Liang et al. 1997, Alshibli and Sture 1999, Sachan and Penumadu 2007). In all cases,
proper lighting conditions were critical for collecting usable photographs for obtaining high
quality results with DIA techniques, as demonstrated by Gachet et al. (2007), Bhandari et al.
(2012), and Hormdee et al. (2014). Examples of the camera-based measurement apparatus are
presented in Figure 2.4.
![Page 33: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/33.jpg)
17
(a) (b)
Figure 2.4. Triaxial testing apparatus with (a) a single, fixed digital camera located outside
of the cell (from Gachet et al. 2007), and (b) three, fixed digital cameras located outside of
the cell (from Bhandari et al. 2012).
2.4.2. Photogrammetric Techniques
All of the previously discussed methods for monitoring soil specimens during triaxial
tests utilized external cameras and therefore several optical challenges were encountered, as 1)
described in detail by Bhandari et al. (2012) and Salazar and Coffman (2015) and 2) as
illustrated in Figure 2.5. Although Kikkawa et al. (2006) first introduced a stereo
photogrammetry technique for measuring local displacement and volume for specimens in
triaxial compression, photogrammetric techniques for monitoring triaxial soil specimens did not
![Page 34: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/34.jpg)
18
reappear in the literature until Salazar et al. (2015) and Zhang et al. (2015). The initial work by
Zhang et al. (2015) was extended by the same authors in Li et al. (2016) and the Salazar et al.
(2015) work was extended in Miramontes (2016) and Salazar et al. (2017a, 2017b). As stated by
Zhang et al. (2015), photogrammetric techniques were needed because of the significant
limitations and assumptions of other photographic methods (e.g. only local volume was
obtainable, or precise control of camera location and orientation were required for accurate
measurements).
Figure 2.5. Optical magnification of soil specimen due to presence of confining fluid.
![Page 35: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/35.jpg)
19
The Zhang et al. (2015) and Li et al. (2016) technique involved acquiring photographs of
the specimen at various angles from outside of the cell wall using a single digital single lens
reflex (DSLR) camera. The images were then used to photogrammetrically construct a three-
dimensional model of the specimen, which was scaled to a real-world coordinate system to
obtain the local and total volume of the specimen for a given stage of testing. The Zhang et al.
(2015) and Li et al. (2016) method presented advantages of a photogrammetric approach by
overcoming many of the limitations of other photograph-based methods. However, the
implementation of the method introduced additional processing complexity by requiring
computationally intensive corrections to account for optical distortions. Specifically, a ray
tracing and least-squares optimization technique were utilized to correct for the light refraction at
the confining water–cell wall and cell wall–atmosphere interfaces, for the cell wall curvature,
and for the deformation of the cell wall under high confining pressures. A schematic illustrating
the photogrammetric principles involved with the Zhang et al. (2015) and Li et al. (2016)
technique is presented as Figure 2.6. The results obtained from the Zhang et al. (2015) and Li et
al. (2016) technique are presented as Figure 2.7.
![Page 36: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/36.jpg)
20
Figure 2.6. (a) Schematic of photogrammetric principles involved in the Zhang et al. (2015)
and Li et al. (2016) methodology, including (b) optical ray tracing, and (c) least-square
estimation (from Li et al. 2016).
Figure 2.7. Representation of specimen deformations obtained during a triaxial test
(modified from Li et al. 2016).
![Page 37: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/37.jpg)
21
Salazar and Coffman (2015, 2016) introduced novel camera instrumentation for
monitoring triaxial specimens during testing. A simple board camera device, modified with a
pinhole aperture (BCPA), was designed and incorporated into a camera system that was placed
inside of the triaxial testing cell. The camera device was designed to allow for immersion within
the confining fluid of the triaxial cell (silicone oil). The full immersion of the device caused the
air space behind the camera aperture to fill with the electronics-grade confining fluid, thereby
overcoming the need for a pressure-resistant housing. In addition to a pressure-resistant housing
being impractical for the high confining pressures reached during a triaxial test (up to 1,035
kPa), a pressure-resistant housing would have required more space than was available within the
triaxial cell. Furthermore, the immersion of the camera parts, including the charge-coupled
device (CCD) sensor within the fluid, allowed for direct observation of the soil specimen within
the triaxial cell. Using this technique, light only traveled through one medium (the confining
fluid). In tandem with the pinhole aperture, the need to account for image distortions introduced
by the differences in the indices of refraction of the various materials (lens, air, oil) was
eliminated. Schematics of the BCPA device and the internal camera system are presented as
Figure 2.8 and Figure 2.9, respectively.
![Page 38: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/38.jpg)
22
Figure 2.8. Schematic of board camera device with pinhole aperture developed at the
University of Arkansas (from Salazar and Coffman 2015).
Figure 2.9. Schematic of internal camera-based photogrammetry system for triaxial testing
developed at the University of Arkansas (modified from Salazar et al. 2015).
![Page 39: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/39.jpg)
23
As reported in Salazar et al. (2015, 2017a, 2017b), the internal camera system was used
to capture photographs of the entire specimen surface during a triaxial test by rotating the camera
towers around the specimen. Just as the Zhang et al. (2015) and Li et al. (2016) photogrammetry
technique relied on ringed automatically detected (RAD) targets, the Salazar et al. (2015, 2017)
technique also relied on RAD targets. In Salazar et al. (2015, 2017a, 2017b), the targets were
printed onto the membrane surrounding the specimen and were used to locate points on the
specimen surface. A close-range photogrammetry technique was then utilized to construct a
three-dimensional model of the specimen. Models that were obtained from various stages of
testing were scaled to a real-world coordinate system and then compared to obtain volumetric
changes. The models also allowed for virtual measurement of specimen dimensions. An example
of the results obtained from the Salazar et al. (2015, 2017a, 2017b) technique is presented as
Figure 2.10.
Figure 2.10. Representation of specimen deformation obtained from close-range
photogrammetry during an undrained, conventional, triaxial compression test (modified
from Salazar et al. 2017b).
![Page 40: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/40.jpg)
24
2.5. References
Adrian, R. J., 1991, "Particle Imaging Techniques for Experimental Fluid Mechanics," Ann. Rev.
Fluid Mech., Vol. 23, No. 1, pp. 261-304.
Alshibli, K. A. and Sture, S., 1999, “Sand Shear Band Thickness Measurements by Digital
Imaging Techniques,” J. Comput. Civ. Eng., Vol. 13, No. 2, pp. 103–109.
Alshibli, K. A. and Al-Hamdan, M., 2001, “Estimating Volume Change of Triaxial Soil
Specimens From Planar Images,” Comput-Aid. Infrastruct. Eng., Vol. 16, No. 6,
pp. 415–421.
ASTM D4767-11, 2011, "Standard Test Methods for Consolidated Undrained
Triaxial Compression Test for Cohesive Soils," ASTM International, West
Conshohocken, PA, www.astm.org. doi:10.1520/D4767-11.
Aversa, S. and Nicotera, M. V., 2002, "A Triaxial and Oedometer Apparatus for Testing
Unsaturated Soils," Geotech. Test. J., Vol. 25, No. 1, pp. 3-15.
Bésuelle, P. and Desrues, J., 2001, “An Internal Instrumentation for Axial and Radial Strain
Measurements in Triaxial Tests,” Geotech. Test J., Vol. 24, No. 2, pp. 193-199.
Clayton, C. R. I. and Khatrush, S. A., 1986, "A New Device for Measuring Local Axial Strains
on Triaxial Specimens," Géotechnique, Vol. 36, No. 4, pp. 593-597.
Clayton, C. R. I., Khatrush, S. A., Bica, A. V. D., and Siddique, A., 1989, “The Use of Hall
Effect Semiconductors in Geotechnical Instrumentation,” Geotech. Test. J., Vol. 12,
No. 1, pp. 69–76.
Cuccovillo, T. and Coop, M. R., 1997, “The Measurement of Local Axial Strains in Triaxial
Tests Using LVDTs,” Géotechnique, Vol. 47, No. 1, pp. 167–171.
Cui, Y.J. and Delage, P., 1996, "Yielding and Plastic Behavior of an Unsaturated Compacted
Silt," Géotechnique, Vol. 46, No. 2, pp. 291-311.
Desrues, J., Chambon, R., Mokni, M., and Mazerolle, F., 1996, “Void Ratio Evolution Inside
Shear Bands in Triaxial Sand Specimens Studied by Computed Tomography.
Géotechnique, Vol. 46, No. 35, pp. 529-546.
Ehrgott, J. Q., 1971, "Calculation of Stress and Strain from Triaxial Test Data no Undrained Soil
Specimens," Miscellaneous Paper S-71-9, U.S. Army Engineer Waterways Experiment
Station, Vicksburg, Mississippi.
Gachet, P., Geiser, F., Laloui, L., and Vulliet, L., 2007, “Automated Digital Image Processing
for Volume Change Measurement in Triaxial Cells,” Geotech. Test. J., Vol. 30, No. 2,
pp. 98–103.
![Page 41: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/41.jpg)
25
Gudehus, G. and Nübel, K., 2004, “Evolution of Shear Bands in Sand,” Géotechnique, Vol. 54,
No. 3, pp. 187–201.
Guler, M., Edil, T. B., and Bosscher, P. J., 1999, Measurement of Particle Movement in Granular
Soils Using Image Analysis," J. Comput. Civ., Vol. 13, No. 2, pp. 116-122.
Hormdee, D., Kaikeerati, N., and Jirawattana, P., 2014, “Application of Image Processing for
Volume Measurement in Multistage Triaxial Tests,” Adv. Mat. Res., Vol. 931-932,
pp. 501-505.
Iskander, M. and Liu, J., 2010, "Spatial Deformation Measurement Using Transparent Soil,"
Geotech. Test. J., Vol. 33, No. 4, 8 pgs.
Jain, S., Wang, Y. H., and Fredlund, D. G., 2015, "Non-Contact Sensing System to Measure
Specimen Volume During Shrinkage Test," Geotech. Test. J., Vol. 38, No. 6,
pp. 936-949. doi:10.1520/GTJ20140274.
Kikkawa, N., Nakata, Y., Hyodo, M., Murata, H., and Nishio, S., 2006, "Three-Dimensional
Measurement of Local Strain Using Digital Stereo Photogrammetry in the Triaxial Test,"
Geomechanics and Geotechnics of Particulate Media, M. Hyodo, H. Murata, and Y.
Nakata, Eds., Taylor and Francis, London, pp. 61-67.
Khan, M. H. and Hoag, D. L., 1979, “A Noncontacting Transducer for Measurement of Lateral
Strains,” Can. Geotech. J., Vol. 16, No. 2, pp. 409–411.
Leong, E. C., Agus, S. S., and Rahardjo, H., 2004, “Volume Change Measurement of Soil
Specimen in Triaxial Test,” Geotech. Test. J., Vol. 27, No. 1, pp. 47–56.
Li, L., Zhang, X., Chen, G., and Lytton, R., 2016, “Measuring Unsaturated Soil Deformations
During Triaxial Testing Using a Photogrammetry-Based Method,” Can. Geotech. J.,
Vol. 53, No. 3, pp. 472-489.
Liang, L., Saada, A., Figueroa, J. L., and Cope, C. T., 1997, “The Use of Digital Image
Processing in Monitoring Shear Band Development,” Geotech. Test. J., Vol. 20, No. 3,
pp. 324–339.
Macari, E. J., Parker, J. K., and Costes, N. C., 1997, “Measurement of Volume Changes in
Triaxial Tests Using Digital Imaging Techniques,” Geotech. Test. J., Vol. 20, No. 1,
pp. 103–109.
Mehdizadeh, A., Disfani, M. M., Evans, R., Arulrajah, A., and Ong, D. E. L., 2016, “Discussion
of ‘Development of an Internal Camera-Based Volume Determination System for
Triaxial Testing’ by S. E. Salazar, A. Barnes and R. A. Coffman
(doi:10.1520/GTJ20140249).” Geotech. Test J., Vol. 39, No. 1, pp. 165-168.
doi:10.1520/GTJ20150153.
![Page 42: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/42.jpg)
26
Messerklinger, S. and Springman, S. M., 2007, “Local Radial Displacement Measurements of
Soil Specimens in a Triaxial Test Apparatus Using Laser Transducers,” Geotech. Test. J.,
Vol. 30, No. 6, pp. 1-12.
Miramontes, L. D., 2016, "Validation of an Internal Camera Based Volume Determination
System for Triaxial Testing," Civil Engineering Undergraduate Honors Theses. 33.
Ng, C. W. W. and Chiu, C. F., 2001, “Behaviour of a Loosely Compacted Unsaturated Volcanic
Soil,” J. Geotech. Geoenviron. Eng., Vol. 127, No. 12, pp. 1027–1036.
Ng, C. W. W., Zhan L. T., and Cui Y. J., 2002, “A New Simple System for Measuring Volume
Changes in Unsaturated Soils,” Can. Geotech. J., Vol. 39, No. 3, pp. 757-764.
Önal, O., Ören, A. H., Özden, G., and Kaya, A., 2008, “Determination of Cylindrical Soil
Specimen Dimensions by Imaging With Applications to Volume Change of Bentonite–
Sand Mixtures,” Geotech. Test. J., Vol. 31, No. 2, pp. 124–131.
Ören, A. H., Önal, O., Özden, G., and Kaya, A., 2006, “Nondestructive Evaluation of
Volumetric Shrinkage of Compacted Mixtures Using Digital Image Analysis,”
Eng. Geol., Vol. 85, No. 3, pp. 239–250.
Pinyol, N. M. and Alvarado, M., 2017, "Novel Analysis for Large Strains Based on Particle
Image Velocimetry," Can. Geotech. J., Vol. 54, No. 7, pp. 933-944. doi:10.1139/cgj-
2016-0327.
Rechenmacher, A. L. and Finno, R. J., 2004, “Digital Image Correlation to Evaluate Shear
Banding in Dilative Sands,” Geotech. Test. J., Vol. 27, No. 1, pp. 13–22.
Rechenmacher, A. L. and Medina-Cetina, Z., 2007, “Calibration of Soil Constitutive Models
with Spatially Varying Parameters,” J. Geotech. Geoenviron., Vol. 133, No. 12,
pp. 1567–1576.
Romero, E., Facio, J. A., Lloret, A., Gens, A., and Alonso, E. E., 1997, “A New Suction and
Temperature Controlled Triaxial Apparatus,” presented at the Fourteenth International
Conference on Soil Mechanics and Foundation Engineering, Hamburg, Germany,
September 6-12, 1997, August Aimé Balkema, Amsterdam, pp. 185-188
Sachan, A. and Penumadu, D., 2007, “Strain Localization in Solid Cylindrical Clay Specimens
Using Digital Image Analysis (DIA) Technique,” Soils Found., Vol. 47, No. 1, pp. 67-78.
Salazar, S. E., Barnes, A., and Coffman, R. A., 2015, “Development of an Internal Camera
Based Volume Determination System for Triaxial Testing,” Geotech. Test. J., Vol. 38,
No. 4, pp. 549-555. doi:10.1520/GTJ20140249.
![Page 43: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/43.jpg)
27
Salazar, S. E., Barnes, A., and Coffman, R. A., 2017a, “Closure to “Discussion of ‘Development
of an Internal Camera-Based Volume Determination System for Triaxial Testing’ S. E.
Salazar, A. Barnes and R. A. Coffman” by A. Mehdizadeh, M. M. Disfani, R. Evans, A.
Arulrajah and D. E. L. Ong,” Geotech. Test. J., Vol. 40, No. 1, pp. 47-51.
doi:10.1520/GTJ20160154.
Salazar, S. E. and Coffman, R. A., 2015a, “Consideration of Internal Board Camera Optics for
Triaxial Testing Applications,” Geotech. Test. J., Vol. 38, No. 1, pp. 40-49.
doi:10.1520/GTJ20140163.
Salazar, S. E. and Coffman, R. A., 2016, “Pressurized Fluid-Submerged, Internal, Close-Range
Photogrammetry System for Laboratory Testing,” U.S. Patent and Trademark Office.
Patent 15/360,820. Filed November, 2016.
Salazar, S. E., Miramontes, L. D., Barnes, A., Bernhardt, M. L., and Coffman, R. A., 2017b,
“Verification of an Internal Close-Range Photogrammetry Approach for Volume
Determination During Triaxial Testing,” Geotech. Test. J. Submitted for Review.
Manuscript Number: GTJ-2017-0125.R2.
Scholey, G. K., Frost, J. D., Lo Presti, C. F., and Jamiolkowski, M., 1995, “A Review of
Instrumentation for Measuring Small Strains During Triaxial Testing of Soil Specimens,”
Geotech. Test. J., Vol. 18, No. 2, pp. 137–156.
Stanier, S. A., Blaber, J., Take, W. A., and White, D. J., 2016, "Improved Image-Based
Deformation Measurement for Geotechnical Applications," Can Geotech. J., Vol. 53,
No. 5, pp. 727-739. doi:10.1139/cgj-2015-0253.
Toyota, H., Sakai, N., and Nishimura, T., 2001, "Effects of Stress History due to Unsaturation
and Drainage Conditions on Shear Properties of Unsaturated Cohesive Soil," Soils
Found., Vol. 41, No. 1, pp. 13-24.
Uchaipichat, A., Khalili, N., and Zargarbashi, S., 2011, “A Temperature Controlled Triaxial
Apparatus for Testing Unsaturated Soils,” Geotech. Test. J., Vol. 34, No. 5, pp. 1-9.
Viggiani, G., Lenoir, N., Bésuelle, P., Di Michiel, M., Marello, S., Desrues, J., and Kretzschmer,
M., 2004, “X-ray Microtomography for Studying Localized Deformation in Fine-Grained
Geomaterials Under Triaxial Compression,” Cr Acad Sci II B-Mec, Vol. 332,
pp. 819-826.
Wheeler, S. J., 1986, "The Stress-Strain Behaviour of Soils Containing Gas Bubbles," Ph.D.
dissertation, Oxford University, Oxford, UK.
White, D., Take, W., and Bolton, M., 2003, "Soil Deformation Measurement Using Particle
Image Velocimetry (PIV) and Photogrammetry," Géotechnique, Vol. 53, No. 7,
pp. 619-631.
![Page 44: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/44.jpg)
28
Zhang, X., Li, L., Chen, G., and Lytton, R., 2015, “A Photogrammetry-Based Method to
Measure Total and Local Volume Changes of Unsaturated Soils During Triaxial
Testing,” Acta Geotechnica, Vol. 10, No. 1, pp. 55-82. doi:10.1007/s11440-014-0346-8.
![Page 45: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/45.jpg)
29
CHAPTER 3: CONSIDERATION OF INTERNAL BOARD CAMERA
OPTICS FOR TRIAXIAL TESTING APPLICATIONS
3.1. Chapter Overview
The concept of optics, internal to a triaxial testing cell, is explored in this chapter. The
challenges, limitations, and advantages associated with this concept are described and a simple
board camera modified with a pinhole aperture (BCPA) is introduced. The challenges of
implementing this camera system included limited space within the testing cell, the presence of
pressurized confining fluid within the cell, and sufficient photographic coverage of the entire
surface of a soil specimen. Preliminary testing of the camera device is presented and a system
that incorporated multiple devices attached to rotating fixtures within the cell are introduced.
The limitations of the included manuscript (Salazar and Coffman 2015) are discussed in
Section 3.2. The full citation for this document is included in Section 3.3. The motivation and
background for the manuscript are described in Sections 3.4, 3.5, and 3.6. Contained within
Section 3.7 are the challenges that were involved with designing a camera system capable of
acquiring images of a specimen from within the testing cell during a triaxial test. Section 3.8
contains a description of the BCPA device that overcame the presented challenges and Section
3.9 includes the testing and calibration of the BCPA device. Conclusions for this work are
presented in Section 3.10.
3.2. Limitations of the Described Study
The BCPA device that was designed for implementation within the internal camera-based
monitoring system suffered from relatively poor image quality, due to the nature of the pinhole
aperture and the low-cost board camera sensor. Although the BCPA overcame the presented
challenges, the optimization of the device required compromises in the field of view, resolution,
and light entry characteristics of the device. Although the study presented the concept of the
![Page 46: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/46.jpg)
30
entire internal camera-based system, the results that were presented were focused primarily on
the development of the BCPA device and not on the system. Therefore, the manuscript presented
the development of the optics that had the potential for determination of specimen volumes.
3.3. Consideration of Internal Board Camera Optics for Triaxial Testing Applications
Reference
Salazar, Sean E. and Coffman, Richard A., “Consideration of Internal Board Camera Optics for
Triaxial Testing Applications,” Geotechnical Testing Journal, Vol. 38, No. 1, 2015, pp. 40-49.
doi:10.1520/GTJ20140163.
3.4. Abstract
The application of small board cameras, located within a triaxial cell to determine radial
and axial strain, was investigated. Specifically, charge-coupled device (CCD) sensors were
utilized in conjunction with precision pinhole apertures to capture images from within the triaxial
cell. The cameras were fully immersed in electronics-grade silicone oil and were able to
withstand cell pressures that are common to triaxial testing (up to 1034 kPa (150 psi)). The small
size of the cameras allowed for implementation within the triaxial cell, thereby avoiding: (1) the
cumbersome corrections that are required to account for refraction at the confining fluid–cell
wall–air interfaces and magnification due to cell wall curvature, and (2) the amount of space
required for outside-of-the-cell monitoring systems that utilize cameras. The final design of
the cameras was based on an iterative testing process in which various types of small board
cameras, lenses, and finally pinhole apertures were investigated. The advantages of the lensless
pinhole aperture camera design included: (1) lack of optical aberrations, such as those
encountered in traditional lensed camera systems, (2) practically infinite depth of field, allowing
for sharp, close-up images, and (3) wide-angle field of view without the distortions that are
associated with the use of wide-angle lenses. As discussed herein, the pinhole cameras were
optimized for optical resolution and light entry to minimize the effect of diffraction patterns that
![Page 47: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/47.jpg)
31
are commonly associated with pinhole apertures. The resolution of the cameras was determined
to be sufficient for the potential application of the cameras (volume measurements). The
instrumentation presented herein provides a novel alternative to the state-of-the-art outside-of-
the-cell photogrammetric instrumentation that is currently employed to monitor soil specimens
during triaxial tests.
Keywords: photogrammetry, refraction, triaxial testing, laboratory equipment
3.5. Introduction
The current state-of-the-art photogrammetric technique for measuring the change in
volume of soil specimens located within a triaxial cell utilizes expensive digital cameras that are
located outside of the triaxial cell. The use of small board cameras located within the triaxial cell,
that overcome pressure and space constraints, has the potential to improve the current state-of-
the-art of photogrammetric techniques for triaxial testing, at a fraction of the expense. The
optical design of the board camera was of particular importance, as it was necessary to develop
high quality images within a confined space, while the camera was immersed in the confining
fluid. Therefore, a small open body board camera with a pinhole aperture (BCPA) was designed.
Although the use of board cameras with traditional lenses was attempted, the utilization of the
BCPA was proven to perform better than the lensed camera and the optical design was simpler,
more feasible, and more cost effective. The disadvantages of using a BCPA include (1) the
need for very precise aperture construction for acquisition of high quality imagery and (2)
limited resolution (when compared to traditional, lensed, outside-of-the-cell cameras). However,
the practical and economic advantages of using a BCPA far outweigh the disadvantages. The
basic optical principles of pinhole apertures and the existing state-of-the-art practice of obtaining
volume measurements using photogrammetry are described and the challenges encountered
![Page 48: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/48.jpg)
32
during the design and fabrication of the BCPA are presented. The images obtained using the
BCPA, located within the triaxial cell, are compared with images obtained from traditional
lensed cameras, located within the triaxial cell, and the differences are discussed. Final remarks
and a summary of the BCPA system are also provided.
3.6. Background
Historically, when photogrammetric techniques were employed to measure the amount of
volume change in specimens being tested in a triaxial device, outside-of-the-cell cameras were
utilized (Parker 1987, Macari et al. 1997, Alshibli and Sture 1999, Alshibli and Al-Hamdan
2001, Gachet et al. 2006, Sachan and Penumadu 2007, Bhandari et al. 2012). However, light
refraction at the (1) confining fluid-cell wall interface and (2) cell wall-atmosphere interface and
the curvature of the cell wall have necessitated the use of models to account for the refraction
and magnification effects. Furthermore, the cameras surrounding the testing apparatus have been
expensive, limited by technology, and have required an excessive amount of space to develop the
required focal length and lighting conditions. Moreover, the optical elements of the camera
equipment have not been addressed in detail, such as optical aberrations, inherent to the camera
lenses.
3.6.1. Lens Optics
Lenses are typically used to capture and focus light and may be used to increase the field
of view. However, errors introduced by refraction of light through lenses, including spherical
aberration, coma, field curvature, astigmatism, and barrel, pincushion, and complex distortions
are prevalent to varying extents in lens applications (Mahajan 1998, Roichman et al. 2006,
Kingslake and Johnson 2010). Most camera lenses are therefore constructed of multiple lenses
(lens array) that are stacked to correct for some aberrations. A careful balance between
![Page 49: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/49.jpg)
33
mitigating one type of aberration and augmenting another type has always existed; therefore, it is
never truly possible to capture an image that does not contain some type of aberration when a
lens or lens array is utilized. Although most cameras use lens arrays, liquid lenses (with variable
focus induced by an electrowetting process) have recently been developed for small applications
(Kuiper and Hendriks 2004, Hendriks et al. 2006, Nguyen 2010) to overcome aberrations
encountered with lenses and to adjust the focal length without the need for mechanical servo
action; liquid lenses are commonly utilized in many smart phone cameras. The aforementioned
focus (inverse of power) of a given lens may be calculated utilizing the Lensmaker’s equation
(Equation 3.1) for thin lenses, first developed by English physicist Thomas Young (1773–1829)
and later by Kuo and Ye (2004):
1
𝑓= (
𝑛1
𝑛2− 1) (
1
𝑟1−
1
𝑟2) Kuo and Ye (2004) Equation 3.1
Where f is the focal length of the lens, n1 is the refractive index of the lens material, n2 is the
refractive index of the surrounding medium, r1 is the radius of curvature of the front surface of
the lens, and r2 is the radius of curvature of the back surface of the lens.
3.6.2. Pinhole Aperture
The pinhole aperture camera is the most basic type of camera and is often overlooked in
favor of a lensed camera. However, despite, and perhaps because of its simplicity, the pinhole
camera may provide: (1) images free of the optical distortions that are inherent to the use of
lenses, (2) images with virtually infinite depth of field, (3) wide viewing angles, and (4) a
foundation for understanding the basic concepts involved in the field of optics, specifically
related to the use of cameras. The primary advantage of using a lens, as opposed to a simple
pinhole, is that a lens can capture and focus more light without requiring long exposure times,
thereby increasing optical resolution (defined as the ability to resolve detail). When resolution is
![Page 50: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/50.jpg)
34
not the most critical objective of a camera application, a simple pinhole aperture camera may
provide a viable alternative to typical lensed camera. Pinholes have been used for centuries for
purposes of viewing and tracing images onto drawings prior to utilizing photo-sensitive materials
for photography purposes (Renner 2000). Moreover, the basic concepts of pinhole optics were
instrumental to the formulation of the theory of light. The theory was supported by the earliest
written observations of multiple phenomena related to light, specifically diffraction, interference,
and polarization of light through pinholes (Grimaldi 1665, Newton 1730, Young 1802, Fresnel
1819). According to Renner (2000), pinholes are still commonly utilized due to the
impracticality of using lenses. For example, scientific application of pinhole imaging may be
found in astro and nuclear physics (such as for high-energy particle imaging of laser plasma, X-
rays, the sun, black holes, and exploding stars).
The design of a pinhole aperture is relatively simple; however, certain considerations are
necessary to optimize image quality. Unlike lensed cameras, pinhole cameras rely on diffraction,
not refraction (Figure 3.1). The theory and equations for pinhole apertures were suggested by
early researchers like Herschel (1835), Airy (1835), and Strutt (1891); attempts have also been
made to refine the relationship between the optical phenomena in more recent years. However, as
Young (1971) indicates, the theoretical limits should only be used as a guideline because the
optimal pinhole aperture diameter is often better determined experimentally. The optimal pinhole
diameter is limited by resolution (larger diameters correspond with poorer resolution), by Fresnel
(near-field) and Fraunhofer (far-field) diffraction limits, and by the ability to gather light (smaller
diameters correspond with higher diffraction interference and allow less light to be collected).
The optimal pinhole diameter for optical applications often relates to the “Airy disk,” which is
the bright, focused spot, central to a diffraction pattern through a perfectly circular aperture.
![Page 51: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/51.jpg)
35
Figure 3.1. Real image formation illustrated by simplified ray diagram of (a) diffraction of
light through a pinhole aperture, and (b) refraction of light through a lens.
3.6.3. Pressure Resistant Cameras
To be able to withstand high pressures, a camera is typically sealed in pressure-resistant
or, more commonly, pressure-compensating housings (Laudo et al. 1998). These housings are
typically bulky, expensive, and do not allow for direct optical observation because light must
first pass through a transparent thermoplastic barrier (i.e., acrylic plastic) before reaching the
camera. To combat this, fluids such as silicone oil or mineral oil (Salazar and Coffman 2014),
may be used in electronics applications where exposure to the fluid is unavoidable or desired.
The direct contact between the electronics and fluid will not cause short-circuiting due to the
inert and non-ionic properties of the oil (Mohapatra and Loikits 2005, Schmidt 2005, Lasance
and Simons 2005). Furthermore, even at high pressures, the silicone oil does not crush the
components of the camera even though the components are directly subjected to the fluid. This
11 11
6. Principle axis
7. Single, thin, convex lens
8. Focal point
9. Distance from object to lens
10. Distance from lens to image plane
11. Distance from object to image plane
9 10
1. Viewing angle
2. Pinhole aperture diameter
3. Aperture substrate thickness
4. Image plane
5. Focal length
5 5
8
8
7 6
4
(a)
1
5
4
2
3
(b)
![Page 52: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/52.jpg)
36
direct immersion allows for pressure resistant design, without the need for a housing; thereby
also allowing for direct optical observation.
3.7. Challenges Encountered
The design of the optical and mechanical components of a camera system that was used
to monitor triaxial specimens from within a triaxial cell, submerged in confining fluid, and
subjected to high pressures is presented herein. The following challenges were encountered and
are addressed sequentially: (1) physical space requirements of placing multiple cameras within a
standard triaxial cell, (2) direct contact between electronic components of the camera and the
confining fluid, (3) space requirements for developing the appropriate focal length, (4) high cell
pressures during testing, (5) sufficient coverage of entire specimen area with minimal camera
deployment, and (6) analog to digital signal conversion for capturing still frames from video
feeds. Discoveries were made through an initial empirical trial and error process, and through
theoretical deductions to test, optimize, and fabricate new BCPA designs. The process that was
followed to address each of the interrelated challenges is presented in Figure 3.2. Specifically, of
most importance was the focal length, as the focal length was a function of all of the other
challenges. Each of the aforementioned challenges is further discussed in the forthcoming
sections.
![Page 53: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/53.jpg)
37
Figure 3.2. The process that was followed to address the interrelated challenges in the
design of the BCPA.
3.7.1. Space Requirements
The first challenge was the small size of the triaxial cell (11.43 cm (4.5 in.)) inside
diameter Trautwein Soil Testing Equipment Co. triaxial cell). Only 3.81 cm (1.5 in.) of space
surrounded the 3.81 cm (1.5 in.) diameter specimen. To overcome this challenge, small closed
circuit board cameras (with dimensions of 14mm (0.55 in.)) by 14mm (0.55 in.) by 13mm (0.51
in.)) were placed into the cell in between the cell wall and the soil specimen. Several types of
cameras were investigated including various types of cameras with a 6.35mm (0.25 in.) format or
8.38mm (0.33 in.) format charge-coupled device (CCD) or complementary metal oxide
semiconductor (CMOS) sensor that were mounted to a circuit board that housed composite video
(yellow), audio (white), and power (red, black) wire leads and enclosed within an aperture box
with a threaded lens mount assembly.
The quality and size of the images produced from the respective cameras was partially a
function of the sensor size; therefore, given the available options, a 8.38mm (0.33 in.) format
sensor was selected over similar 6.35mm (0.25 in.) format sensors. Each of the cameras was
tested with a variety of standard lenses both in air and submerged in electronics-grade silicone
Space requirements (SR)
Confining fluid (CF)
Focal length (FL)
Cell pressure (CP)
Specimen coverage (SC)
Challenges Alternative Pursued
Small board cameras
Further challenges
PSF-5cSt silicone oil
Various lenses
Multiple cameras
Open body camera
Limited field of view
Refractive index of oil,
compressible air spaces
Solution
14x14x13mm (SR, SC)
Open body (CF, CP)
8 cameras (SC)
Rotating towers (SC)
BCPA
![Page 54: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/54.jpg)
38
oil. Furthermore, only cameras with the highest lines of horizontal resolution (LoHR), and pixel
dimensions were selected (specifications ranged from 420 LoHR to 700 LoHR and 492 by 510 to
976 by 582, respectively). The board camera that was selected for use in the triaxial cell
possessed a 8.38mm (0.33 in.) format SONY CCD sensor, capable of obtaining 700 LoHR, and
976 horizontal by 582 vertical effective pixels. This camera was chosen because it had the best
image quality to size ratio, thereby facilitating deployment inside of the triaxial cell. Photographs
of the various types of cameras and lenses that were tested, and their respective specifications,
are presented in Figure 3.3.
Given the space constraints within the triaxial cell, it was not practical to design a lens
array to focus light for this application. It may have been possible to design a lens system to
improve the resolution of captured images; however, given the space constraints, the design
would have required using very small lenses (on the order of 2.0mm in diameter) with unusually
high refractive indices (greater than 1.8) to give the appropriate focal length when immersed in
the silicone oil. Specifically, the appropriate focal length that was required is discussed in the
section entitled focal length.
![Page 55: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/55.jpg)
39
Figure 3.3. (a) Three types of board cameras that were tested, and (b) five different types of
lenses used with the cameras.
3.7.2. Confining Fluid
To avoid damage to the sensitive electronics of the cameras, while still ensuring
saturation of the specimen (by utilizing pressurized fluid instead of pressurized air to prevent gas
diffusion across the membrane), silicone fluid (PSF-5cSt, Trautwein Soil Testing Equipment
Co.) was used to confine the specimens and surrounded the cameras. Properties of the silicone
fluid include: low viscosity (5cSt), specific gravity of 0.918, dielectric constant of 2.60, dielectric
strength of 375, and index of refraction of 1.397. Due to the high refractive index of the oil
(relative to air), the standard lenses were not able to focus when immersed in the oil.
Specifically, the index of refraction of the silicone oil (1.397) was much greater than the index of
A
B
C
(a)
(b)
1 2 3
4
5
Board camera specifications:
A) 6.35mm [0.25in.] format Sharp
CCD sensor, 420 LoHR, 492×510,
M6.5×0.25 aperture box threading,
B) 8.38mm [0.33in.] format Sony
CCD sensor, 700 LoHR, 976×494,
M12×0.5 aperture box threading;
C) 8.38mm [0.33in.] format Sony
CCD sensor, 700 LoHR, 976×582,
M12×0.5 aperture box threading.
Lens specifications:
1) 3.6mm standard lens, 55° FOV;
2) 3.7mm button lens, 60° FOV;
3) 2.8mm barrel lens, 90° FOV;
4) 2.1mm wide angle lens, 170° FOV;
5) barrel-mounted pinhole aperture.
Note: All had M12×0.5 threading but
Lens 1 (M6.5×0.25 thread).
![Page 56: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/56.jpg)
40
refraction of air (1.000 in a perfect vacuum). Although unknown, it was estimated that the index
of refraction of the lens material was between 1.48 and 1.60 (for crown or flint glass). The
increase in the index of refraction from 1.000 to 1.397 reduced the difference in the indices of
refraction between the two media (air–glass and oil–glass) and thereby increased the required
focal length. This reduction in the difference in the indices of refraction provided for non-ideal
light dispersion and therefore led to severely out-of focus images when the cameras containing
lenses were submerged in oil. Simply put, lenses were deemed to not be a viable option (as
discussed in further detail in the focal length section).
3.7.3. Focal Length
As discussed previously, limited space was available to deploy the cameras. This space
constraint limited the maximum achievable focal length (the distance between the object and the
lens, as shown in the ray diagram that was previously presented in Figure 3.1). The minimum
focal length values, for the standard lenses that were included with purchase of the board
cameras when tested in air, were between 4 and 6 cm (1.6 and 2.4 in.). These distances
corresponded to images with the highest sharpness; therefore, focused images could not be
obtained when using cameras with the standard lenses within the available confined space of
3.81 cm (1.5 in.). Furthermore, as revealed by submerging the cameras in the silicone oil, the
focal length of a given lens varied, depending on the medium that surrounded the lens.
Simply put, a camera lens designed to provide focus in air, did not provide focus in the
silicone oil. Although different types of lenses were tested (Figure 3.3), all of the lenses inhibited
viewing of soil specimens when the lenses were immersed in silicone oil. The testing of the
lenses was purely empirical, due to the unknown characteristics of the various lenses (lens shape,
refractive index of lens material, and radii of curvature of the lens surfaces). The characteristics
![Page 57: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/57.jpg)
41
of the lenses were unknown because the small lenses were cemented and sealed inside of a
mounting assembly, making it impractical to extract the lenses or lens components for closer
inspection. The Lensmaker’s equation (previously presented as Equation 3.1) was employed to
determine the optical properties of a lens that would enable collection of images when immersed
in silicone oil. However, given the immersion medium, it was not economically, nor practically
feasible to purchase or (manufacture) a lens, or lens array, with the correct refractive index
(greater than 1.8) to develop the appropriate focal length (approximately 24mm (0.94 in.)) within
the physical space limitations (3.81 cm (1.5 in.)). Therefore, as discussed in the section entitled
Pinhole Solution, another solution was realized to enable collection of images from close
distances within a fluid with a high refractive index.
3.7.4. Cell Pressure
To withstand the cell pressures during testing (up to 1034 kPa (150 psi)), the cameras
were flooded behind the aperture, filling all of the air space with oil. It was observed, in original
testing of the lensed cameras, that the focal length of the lens arrays permanently changed after
being subjected to typical pressures. This was attributed to the compression of the small void
spaces between the lenses when subjected to pressure, resulting in permanent deformation of the
lenses and thereby altering the optical properties of the lens array. It was therefore determined
that an alternative to a lens array must be developed to withstand pressure applications.
3.7.5. Specimen Coverage
Due to the close proximity of the board camera to the soil specimen, it was not possible
to observe the entirety of the specimen with a single camera. Specifically, the field of view of an
individual camera (21mm (0.83 in.)) was smaller than the height of the specimen (7.62 cm (3.0
in.)). Therefore, by using multiple cameras, individual areas of the specimen were monitored and
![Page 58: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/58.jpg)
42
the photographs of the individual areas were stitched together using post-processing software. To
achieve this, a camera monitoring system was designed with two arrays of four BCPA (eight
total cameras). The BCPAs were mounted to towers that were rotated along a track that was
attached to the base inside of the cell. The track was designed to rotate using pairs of small
magnets; one magnet was mounted to the track and the other magnet was located at various
positions outside of the triaxial cell wall. The use of magnets allowed for the BCPAs to capture
still frames at prescribed intervals during the rotation. A schematic of the track and camera tower
system is presented in Figure 3.4.
Figure 3.4. Schematic of guided camera track system mounted on triaxial apparatus base.
6. Camera cable
7. Rotating Delrin® bearing track
8. Delrin® bearing base
9. Magnet
10. Triaxial apparatus base
8
9
3
4
7
1. BCPA (typical, 8 quantity)
2. Soil sample
3. Camera tower
4. Acrylic top cap
5. Drain lines
1
2
5
5
10
6
5
![Page 59: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/59.jpg)
43
In the original design of the cameras, it was hypothesized that using wide-angle lenses
would be sufficient to capture large areas of the specimen and, therefore, very few cameras
would be required. However, the use of this type of lens was not permitted (due to the
aforementioned problems associated with lenses not enabling image collection when submerged
in silicon oil, due to the physical size requirements of the lenses, and due to the distortions that
were associated with wide-angle lenses (barrel, pincushion, and complex distortions)). Although
these distortions are now a moot point because the wide-angle lens could not be used, these
distortions are known to be difficult to correct and typically reduce the size of the image (due to
cropping requirements).
3.8. Pinhole Solution
To overcome the limitations of focal length, refractive properties of the confining fluid,
cell pressure, and specimen coverage, a lensless pinhole aperture was developed. The required
aperture diameter was approximated, based on Equation 3.2 (Strutt 1891):
𝑓 = 2𝑟2/𝜆 Strutt (1891) Equation 3.2
Where f is the focal length, r is the radius of the pinhole opening (or aperture), and λ is the
design wavelength.
However, unlike for a lens, the f variable used in Equation 3.2 was associated with the
distance between the pinhole aperture and the camera sensor plane, as previously depicted in
Figure 3.1. Furthermore, as discussed previously, unlike lensed cameras, pinhole cameras have
infinite depth of field; therefore, this type of aperture allowed for the entire image to be in focus.
To obtain sharp images and to maximize resolution, the edges of the pinhole must be precisely
cut and the diameter of the pinhole must be small. Moreover, the thickness of the substrate must
be thin to allow for the widest viewing angle (as shown previously in Figure 3.1). Therefore,
![Page 60: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/60.jpg)
44
various pinhole sizes 75, 100, and 150 μm (2.95×10-3, 3.94×10-3, and 5.91×10-3 in.) were laser
cut into the center of a 9.5mm (0.375 in.) diameter wafer substrate (National Aperture, Part
Number 1-75+ B-2, 1-100+ B-2, and 1-150+ B-2), respectively. The steel substrate (300 series
stainless steel) had a thickness of 12.7 μm (5×10-4 in.) and both sides were blackened (+B-2) to
absorb any stray light within the aperture box.
The design optical wavelength (415 nm) was selected based on the results obtained from
a relative light intensity test that was conducted by examining a diffuse reflectance
fluoropolymer reference material (Spectralon, Labsphere, Inc.) using a spectroradiometer (ASD
FieldSpec Pro HandHeld 2 portable spectroradiometer). Although a peak value of 580nm was
observed, a reduced value of 415nm was utilized because of the refractive index ratio (1.397)
that was associated with silicone oil being used as the confining fluid instead of air. Furthermore,
this wavelength (415 nm) was selected because the final position of the pinhole aperture was
fine-tuned (in relation to the camera image plane) using a threaded barrel that screwed into the
aperture box.
A recess was placed into the threaded barrel to enable the wafer substrate to be mounted
to the barrel. The aforementioned three aperture diameters were tested at various distances from
the camera sensor, and it was found that the 75 μm (3.94×10-3 in.) diameter aperture provided
the best image quality when the substrate was located approximately 3.0mm away from the
image plane, as assessed by visual inspection of the acquired images. Therefore, the final design
components of the BCPAs are thus: (1) an 8.38mm (0.33 in.) format board camera encased in an
aperture box with a M12×0.5 threaded opening; (2) a threaded barrel that was used for seating
and adjusting the pinhole aperture substrate; and (3) a laser cut pinhole aperture (75 μm
opening), centered at a specified focal length (3.0 mm) from the camera sensor. A schematic and
![Page 61: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/61.jpg)
45
4. Aperture box
5. Video signal and power supply cable
6. CCD sensor (5.0mm × 4.8mm)
1. Pinhole aperture (75μm diameter)
2. Pinhole substrate (9.5mm diameter, 12.7μm thick)
3. Threaded barrel (M12×0.5, 6.5mm inside diameter)
1. Substrate (9.5mm diameter) with pinhole (75μm
diameter)
2. Threaded barrel (M12×0.5, 6.5mm inside diameter)
3. Aperture box
4. Video signal and power supply cable
a photograph of the assembled BCPA are presented in Figures 3.5 and 3.6, respectively. As
discussed in the section entitled images collected, the BPCA design and fabrication enabled
images to be collected from inside of the triaxial cell while the cameras were immersed in
silicone oil and subjected to high pressures.
Figure 3.5. Schematic of (a) front view, (b) exploded side view, and (c) exploded orthogonal
view of the BCPA.
Figure 3.6. Photograph of one of the BCPAs utilized inside of the triaxial cell.
3.0 mm
9.5 mm
3.0 mm
6.5 mm
14.0 mm
(a)
1
(b) (c)
14
.0 m
m
13.0
mm
2
2
1.0 mm
3
4
5
6
5
6
1
2
4 3
![Page 62: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/62.jpg)
46
3.8.1. Video Signal Acquisition
The video cables of the cameras were connected to a wire harness that was connected to a
nine-pin feedthrough connector located within the top cap of the triaxial device. The feedthrough
allowed for electrical signals to travel into and out of the triaxial device. The video wires that
were connected to the cameras were also connected to the pins on the nine-pin feedthrough
connector; the opposite sides of the pins were connected to the input channels of an eight-way
video/audio switch (Maituo MT-VIKI 8 Port VGA Switch). The single video output channel
from the video/audio switch was then connected to a Universal Serial Bus 2.0 Digital Video
Adapter (Sabrent USB-AVCPT). Each camera was supplied with external power (DC 12V) from
a common external power supply (Enercell 3-12VDC 1A AC Adapter) that provided power to all
of the cameras simultaneously via the nine-pin feedthrough connector. The video feed from each
of the cameras was subsequently received and displayed by switching the video/audio switch.
The software that was included with the video adapter (Sabrent USB 2.0 Video Capture Creator
with Audio) was utilized to capture still frames from the video feed that was obtained from each
of the cameras.
3.9. Images Collected
Still frames, captured from the video feed of a board camera with a lens (located in air
and immersed within PSF-5cSt silicone oil), are presented in Figures 3.7(a) and 3.7(b),
respectively. Still frames, captured with the BCPA (located in air and immersed within PSF-5cSt
silicone oil), are presented in Figures 3.7(c) and 5.7(d), respectively. An example of linear
distortion in images captured using a lens is evident in Figure 3.7(a) and the inability of the
camera to collect focused light through the lens to form a real image is displayed in Figure
3.7(b). As explained previously, because the lens was designed to work in air, the index of
![Page 63: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/63.jpg)
47
refraction of the silicone oil prevented the camera that was fully immersed within oil from
obtaining a focused image.
Figure 3.7. Still frames captured using a 8.38mm [0.33in.] format CCD board camera with
1) 3.7mm button lens (55° FOV, M12×0.5 thread) in (a) air, and within (b) PSF-5cSt
silicone oil, and with 2) 75μm diameter pinhole aperture (attached to a M12×0.5 barrel
with 6.5mm diameter opening) in (c) air, and within (d) PSF-5cSt silicone oil.
It was determined that the captured light on the far left and far right edges of the images
collected using the BCPA faded abruptly and completely (as indicated by the areas to the left and
right of the dashed white lines in Figures 3.7(c) and 3.7(d)). This phenomenon was attributed to a
combination of physical and optical influences. The aperture barrel material blocked the edges of
the CCD sensor along the longer (horizontal) side of the CCD sensor (due to the proximity of the
barrel to the sensor). Thereby, light was prevented from reaching the edges of the CCD sensor
that led to the black appearance. Although not required, the proximity limitation should be
(a) (c)
(d) (b)
![Page 64: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/64.jpg)
48
overcome by enlarging the inside diameter of the threaded barrel. However, it was determined
that, given the experimental equipment, the dimensions of the shorter (vertical) side of the sensor
provided sufficient coverage of the object (if the camera was rotated in such a way that the
camera cable exited from the camera in the horizontal plane as shown previously in Figure 3.4)
and therefore an increase in the inside diameter of the barrel was not necessary. Furthermore, the
“airy disk” covered the entire camera sensor so no visible diffraction patterns were present.
As observed in the comparison between images captured with a lens and those captured
with a pinhole aperture, there was a significant difference in the amount of light exposure. The
lens (Figures 3.7(a) and 3.7(b)) allowed for maximum light gathering. The pinhole aperture
(Figures 3.7(c) and 3.7(d)) allowed for minimal light entry, due to the small size of the opening
(75 μm (3.94×10-3 in.) diameter). In typical pinhole photography, this minimal amount of light
entry is commonly overcome with longer exposure times; however, for the type of board camera
that was used, it was not possible to control the exact exposure time (electronic shutter time
varied between 1/60 and 1/100 000 s, as per the camera manufacturer). Furthermore, the board
camera switched into “night mode” (monochromatic light gathering) when the illuminance levels
dropped below a certain threshold (0.1 lux). Chromatic aberration may have been present in the
captured images, but it was not possible to detect this type of aberration due to the
monochromatic nature of the images. Although the lighting was not modified to collect the
images presented in Figure 3.7, it is recommended that the lighting surrounding the soil sample
be enhanced and controlled to aid in collection of higher quality imagery. Specifically, utilizing
two 17.8 cm (7 in.) diameter dome light sources to surround the entire triaxial chamber will
enhance the imagery. With the aid of the guided camera track system, multiple still frames were
captured with the individual BCPAs along the length of the soil specimen, at prescribed intervals
![Page 65: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/65.jpg)
49
during rotation around the circumference of the specimen. Because of the overlapping fields of
view of adjacent BCPAs, in the vertical direction and in circumferential direction, common
points were acquired within captured images, and the individual geopositions of the cameras
were calculated, allowing for post-process stitching of the collected images. PhotoModeler (Eos
Systems, Inc. 2014) was utilized to calculate the photogrammetric properties of the BCPA (Table
3.1). These properties included the focal length, format size (physical dimensions of the sensor)
and principal point (intersection between principal axis and image sensor). PhotoModeler was
also used to determine the geoposition of each of the individual BCPAs. Specifically, the
positions of the BCPAs were determined by using unique, pre-selected targets that were adhered
to a 1.5 in. (38mm) diameter by 3 in. (76mm) tall brass specimen and that were automatically
recognized within the software. The PhotoModeler obtained photogrammetric properties and
geopositions corresponded well with manual (caliper) measurements.
Table 3.1. Photogrammetric properties of the BCPA.
Focal length (mm) Format size (mm) Principal point (pixels)
3.50 5.16 width 2.62 x
4.80 height 2.23 y
Using repeatable rotation intervals, and therefore known geopositions of each of the
individual BCPAs, PhotoModeler was used to match common points within the captured images
that thus enabled generation of a point cloud for any object that was viewed by the BCPAs. This
point cloud was then meshed to calculate the dimensions and volume of the viewed object. By
utilizing this experimental method (PhotoModeler), the volume of the brass specimen that was
obtained was 91.92 cm3. The volume obtained using manual (caliper, pi tape) measurements was
91.93 cm3, resulting in an estimated error of 0.01 %.
![Page 66: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/66.jpg)
50
3.10. Conclusions
A small board camera with a pinhole aperture was designed for deployment inside of a
triaxial cell to enable measurement of the volume of soil specimens. Because the camera
components were fully immersed in oil and were located very close to the soil specimen, special
considerations were accounted for when designing the optical components of the
photogrammetric instrumentation. To resist the high pressures that are commonly encountered
within the triaxial cell during triaxial testing (up to 1034 kPa (150 psi)), the silicone oil was
allowed to enter behind the camera face and to surround the CCD sensor. The relatively high
refractive index of silicone oil (as compared to the refractive index of air) influenced the light
entering into the traditional lenses or lens arrays yielding severely out of focus images (because
the refractive index of the lens or lens array closely matched the refractive index of the lens
confining fluid).
Utilizing the Lensmaker’s equation, it would have only been possible to focus an image
using small lenses with unusually high refractive indices. However, this option was not pursued
because it was (1) limited by availability, (2) costprohibitive, and (3) required the design and
fabrication of additional lens mounts and boxes. Instead, the as provided lens that was located
within the aperture box of each of the board cameras was replaced with a newly created high-
precision pinhole aperture. The pinhole cameras were designed, fabricated, tested, and their
potential applicability inside of a triaxial cell evaluated. Specifically, high quality images were
acquired using the BCPA even when the BCPA was placed inside of the triaxial cell, immersed
in silicone oil, and subjected to high pressures. Furthermore, a guided track system was designed
to allow for coverage of the entire soil specimen, while deploying the minimal number of BCPAs
within the triaxial cell.
![Page 67: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/67.jpg)
51
3.10.1. Advantages and Limitations of a BCPA
In summary, a careful balance existed between resolution, light entry, and field of view.
To truly optimize the design of the camera, it was necessary to experiment with a variety of
different pinhole diameters and focal lengths. There are many advantages to BCPAs, two of
which are an infinite depth of field (including very close depths) and a lack of any of the optical
aberrations associated with lenses. Other advantages of using a pinhole-type camera are as
follows: the BPCA: (1) is not adversely affected by the refractive properties of the immersion
fluid (silicone oil); (2) requires very little space to develop appropriate focal length and requires
less space than a lens or lens array; (3) may be designed to provide very large viewing angles
without the need for a lens; and (4) can withstand pressure. The disadvantages are primarily low
light entry and limited resolution; however, with proper design and fabrication, these
disadvantages were overcome. By utilizing a BCPA, images were obtained within a triaxial cell
even though the optical design of the BCPA was simplified from that of a traditional lensed
camera.
![Page 68: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/68.jpg)
52
3.11. References
Airy, G. B., 1835, “On the Diffraction of an Object-Glass With Circular Aperture,” Trans.
Cambr. Philos. Soc., Vol. 5, pp. 283–291.
Alshibli, K. A. and Sture, S., 1999, “Sand Shear Band Thickness Measurements by Digital
Imaging Techniques,” J. Comput. Civ. Eng., Vol. 13, No. 2, pp. 103–109.
Alshibli, K. A. and Al-Hamdan, M., 2001, “Estimating Volume Change of Triaxial Soil
Specimens From Planar Images,” Comput-Aid. Infrastruct. Eng., Vol. 16, No. 6,
pp. 415–421.
Bhandari, A. R., Powrie, W., and Harkness, R. M., 2012, “A Digital Image-Based Deformation
Measurement System for Triaxial Tests,” Geotech. Test. J., Vol. 35, No. 2, pp. 1–18.
Eos Systems Inc., 2014, PhotoModeler (Version 2014.1.1). 3D Measurements and Models
Software, http://www.photomodeler.com/index.html (Last accessed 13 Oct 2014).
Fresnel, A.-J., 1819, “Memoir on the Diffraction of Light,” The Wave Theory of Light, H. Crew,
Trans., American Book Company, New York.
Gachet, P., Geiser, F., Laloui, L., and Vulliet, L., 2006, “Automated Digital Image Processing
for Volume Change Measurement in Triaxial Cells,” Geotech. Test. J., Vol. 30, No. 2,
pp. 98–103.
Grimaldi, F. M., 1665, Physico–mathesis de lumine, coloribus, et iride, aliisque adnexis libri duo
[A physiomathematical thesis on light, colors, the rainbow and other related topics in two
books], Bologna, Italy (in Latin).
Hendriks, B. H. W., Kuiper, S., van As, M. A. J., Renders, C. A., and Tukker, T. W., 2006,
“Variable Liquid Lenses for Electronic Products,” Proc. SPIE, Vol. 6034, pp. 10–18.
Herschel, J. F. W., 1835, Treatise on Astronomy, 3rd ed., Carey, Lea & Blanchard,
Philadelphia, PA.
Kingslake, R. and Johnson, R. B., 2010, Lens Design Fundamentals, 2nd ed., Academic Press,
New York.
Kuiper, S. and Hendriks, B. H. W., 2004, “Variable-Focus Liquid Lens for Miniature Cameras,”
Appl. Phys. Lett., Vol. 85, No. 7, pp. 1128–1130.
Kuo, C.-H. and Ye, Z., 2004, “Sonic Crystal Lenses that Obey the Lensmaker’s Formula,”
J. Phys. D., Vol. 37, No. 15, pp. 2155–2159.
Lasance, C. and Simons, R., 2005, “Advances in High Performance Cooling for Electronics,”
Electron. Cool., Vol. 11, No. 4, pp. 22–39.
![Page 69: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/69.jpg)
53
Laudo, J. S., Wurm, K., and Dodson, C., 1998, “Liquid-Filled Underwater Camera Lens
System,” Proc. SPIE, Vol. 3482, pp. 698–702.
Macari, E. J., Parker, J. K., and Costes, N. C., 1997, “Measurement of Volume Changes in
Triaxial Tests Using Digital Imaging Techniques,” Geotech. Test. J., Vol. 20, No. 1,
pp. 103–109.
Mahajan, V. N., 1998, Optical Imaging and Aberrations: Part I. Ray Geometrical Optics, SPIE,
Bellingham, WA.
Mohapatra, S. C. and Loikits, D., 2005, “Advances in Liquid Coolant Technologies for
Electronics Cooling,” Proceedings of the 21st IEEE Semiconductor Thermal
Measurement and Management Symposium, San Jose, CA, March 15–17, pp. 354–360.
Newton, I., 1730, Opticks: Or a Treatise of the Reflections, Refractions, Inflections and Colours
of Light, 4th ed., William Innys, London.
Nguyen, N.-T., 2010, “Micro-Optofluidic Lenses: A Review,” Biomicrofluidics, Vol. 4, No. 3,
031501.
Parker, J. K., 1987, “Image Processing and Analysis for the Mechanics of Granular Materials
Experiment,” ASME Proceedings of the 19th SE Symposium on System Theory, ASME,
New York, pp. 536–541.
Renner, E., 2000, Pinhole Photography: Rediscovering a Historic Technique, 2nd ed.,
Butterworth–Heinemann, Boston, MA.
Roichman, Y., Waldron, A., Gardel, E., and Grier, D. G., 2006, “Optical Traps with Geometric
Aberrations,” Appl. Opt., Vol. 45, No. 15, pp. 3425–3429.
Sachan, A. and Penumadu, D., 2007, “Strain Localization in Solid Cylindrical Clay Specimens
Using Digital Image Analysis (DIA) Technique,” Soils Found., Vol. 47, No. 1,
pp. 67–78.
Salazar, S. E. and Coffman, R. A., 2014, “Design and Fabrication of End Platens for Acquisition
of Small-Strain Piezoelectric Measurements During Large-Strain Triaxial Extension and
Triaxial Compression Testing,” Geotech. Test. J., Vol. 37, No. 6, pp. 948–958.
Schmidt, R., 2005, “Liquid Cooling is Back,” Electron. Cool., Vol. 11, No. 3, pp. 34–38.
Strutt, J. W., Lord Rayleigh, 1891, “On Pin-Hole Photography,” Philos. Mag., Vol. 31,
pp. 87–99.
Young, T., 1802, “Bakerian Lecture: On the Theory of Light and Colours,” Philos. Trans. R.
Soc. Lond., Vol. 92, pp. 12–48.
![Page 70: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/70.jpg)
54
Young, M., 1971, “Pinhole Optics,” Appl. Opt., Vol. 10, No. 12, pp. 2763–2767.
![Page 71: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/71.jpg)
55
CHAPTER 4: DEVELOPMENT OF AN INTERNAL CAMERA-BASED VOLUME
DETERMINATION SYSTEM FOR TRIAXIAL TESTING
4.1. Chapter Overview
The development of the internal camera-based volume determination system is described
in this chapter. The instrumentation presented within this manuscript provided a novel alternative
to the state-of-the-art of camera-based monitoring of triaxial tests. The individual components of
the system, the photogrammetric methodology, and preliminary testing of the system are
detailed.
The limitations of the included manuscript (Salazar et al. 2015) are discussed in
Section 4.2. The full citation for this document is included in Section 4.3. The motivation and
background for the manuscript are described in Sections 4.4, 4.5, and 4.6. Contained within
Section 4.7 is a detailed description of the camera system, including the mechanical and
electrical components of the system (Sections 4.7.1 and 4.7.2, respectively). The
photogrammetric methodology and early results are detailed in Section 4.8. Finally, conclusions
for this manuscript are presented in Section 4.9.
4.2. Limitations of the Described Study
The manuscript contained within this chapter was originally published as a technical note
in order to follow up the Salazar and Coffman (2015) publication that first introduced the
internal camera-based volume determination system. The length of the manuscript was therefore
limited. Although the preliminary testing of the system was described as a "validation process",
the tests did not include any triaxial tests on soil specimens, nor were the camera-based
measurements subject to immersion in confining fluid (all tests were performed in an air-filled
cell only). Furthermore, the photogrammetric methodology that was utilized to determine the
![Page 72: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/72.jpg)
56
volume of a dummy specimen within the testing cell was described in broad terms, but an
overview of the implementation of the system was provided.
4.3. Development of an Internal Camera-Based Volume Determination System for Triaxial
Testing
Reference
Salazar, Sean E., Barnes, Adam, and Coffman, Richard A., “Development of an Internal
Camera-Based Volume Determination System for Triaxial Testing,” Geotechnical Testing
Journal, Vol. 38, No. 4, 2015, pp. 549-555. doi:10.1520/GTJ20140249.
4.4. Abstract
A triaxial testing cell was instrumented with an internal camera monitoring system. By
placing the camera monitoring system inside of the triaxial cell, optical distortions due to
refraction at the confining fluid–cell wall and cell wall–atmosphere interfaces and the
curvature of the cell wall were eliminated. The components of the system are presented.
Furthermore, the photogrammetric techniques that were utilized to analyze the
photographs that were captured from within the triaxial cell are discussed. The proposed
methods for acquiring and analyzing the photographs are presented and the potential for
the inclusion of an internal camera–monitoring system for triaxial testing applications are
discussed.
Keywords: triaxial testing, laboratory equipment, photogrammetry
4.5. Introduction
Of the unconventional testing methods (photogrammetry, other digital imaging
techniques, proximity sensors, x-ray-computed tomography) used to monitor saturated and
unsaturated soil specimens during triaxial testing, photograph-based measurement is a practical,
cost-effective, and versatile method. Photogrammetry may be utilized to: (1) characterize the
failure plane within a soil specimen during testing, (2) monitor the critical cross-sectional area of
![Page 73: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/73.jpg)
57
the soil specimen (bulging or necking behavior), (3) calculate the volume of the soil specimen,
and (4) calculate the volumetric strain within the soil specimen. Several drawbacks exist with
current photograph-based instrumentation, namely the optical effects caused by the curvature of
the cell wall, refraction at the confining fluid–cell wall and cell wall–atmosphere interfaces, and
optical distortions inherent to lensed cameras. These drawbacks must be overcome and corrected
using cumbersome models, further complicating the procedure of acquiring and processing
images.
As described in Salazar and Coffman (2015), the optical components of internal
photogrammetric instrumentation (cameras located within the cell fluid on the inside of the
triaxial cell) for triaxial testing applications were designed, fabricated, and tested to overcome
the aforementioned drawbacks, as well as to overcome the challenges of internal instrumentation
(space, confining fluid, focal length, cell pressure, and specimen coverage). Details about the
photogrammetric system that was placed within the triaxial cell to allow for direct, unobstructed
observation of a specimen during triaxial testing are presented herein. The system allowed for
viewing of the entire specimen surface in both the axial and radial directions. Calibration and
validation of the system was attained by utilizing a photogrammetric technique to digitally
reconstruct the exterior shape of a specimen with known dimensions. The methodology that was
employed to reconstruct the exterior surface and the accuracy and precision of the
photogrammetric measurements are presented and discussed for completeness.
4.6. Background
Specimen volume and volumetric strain measurements have historically been calculated
for soil specimens, during triaxial testing, by utilizing pore fluid volume measurements (Bishop
and Donald 1961, Ng et al. 2002, Leong et al. 2004). These pore fluid measurements have
![Page 74: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/74.jpg)
58
typically been supplemented with data obtained from axial deformation measurements to obtain
the average specimen dimensions, during or after testing, by adding or subtracting the
measurements from the dimensions of the specimen that were manually measured (caliper and pi
tape) prior to testing. Because the volume measurements have been attained by measuring the
change in the amount of pore fluid, the measurements have been affected by temperature- and
pressure-induced flexure of the cell wall and drain lines. Therefore, only estimates, not exact
values, of volumetric strain have been obtained from these measurements. Likewise,
conventional axial (Scholey et al. 1995, Cuccovillo and Coop 1997, Ng and Chiu 2001) and
lateral or radial (Khan and Hoag 1979, Clayton et al. 1989, Bésuelle and Desrues 2001)
measurements also rely on averaging methods that have not accurately accounted for irregular
surfaces. Furthermore, past measurements have been limited to global volume changes that
prevented the characterization of local strains during the development of shear bands, bulging
bifurcation, or, in the case of extension testing, necking.
To overcome these limitations, digital imaging techniques, including digital image
analysis (DIA), digital image correlation (DIC), and particle image velocimetry (PIV), have been
used to monitor deformations within soil specimens by using external (cameras located outside
of the triaxial cell) cameras. Specifically, these techniques have been of increasing interest as an
alternative method to calculate the volumetric strain of soil specimens (Macari et al. 1997,
Alshibli and Al-Hamdan 2001, Puppala et al. 2004, Rechenmacher and Finno 2004, Ören et al.
2006, Gachet et al. 2007, Sachan and Penumadu 2007, Önal et al. 2008, Bhandari et al. 2012)
and to monitor the evolution of shear banding and strain localization (Alshibli and Sture 1999,
Nübel and Weitbrecht 2002, Gudehus and Nübel 2004, Rechenmacher and Medina-Cetina 2007,
Sachan and Penumadu 2007). Photograph-based (DIA, DIC, PIV, photogrammetry)
![Page 75: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/75.jpg)
59
measurements have been shown to correlate well with conventional volume measurements of the
soil specimen within the triaxial apparatus during testing (Macari et al. 1997, Alshibli and Sture
1999, Alshibli and Al-Hamdan 2001, Gachet et al. 2007, Rechenmacher and Medina-Cetina
2007, Sachan and Penumadu 2007, Bhandari et al. 2012, Zhang et al. 2015).
Video feeds and/or still frames of the soil specimens, within the triaxial device, were
captured with external photographic equipment in all of the aforementioned photograph-based
measurement studies. For volumetric measurements, the external instrumentation allowed for
capture of the entire length of the specimen (axial dimension) within a single image; however,
various methods were employed to capture the entire surface area (lateral dimension) of the
specimen. These methods included the use of multiple cameras placed at intervals around the
outside of the cell (Alshibli and Al-Hamdan 2001, Bhandari et al. 2012). In some instances
(Macari et al. 1997, Gachet et al. 2007), the entire specimen surface was not captured and,
therefore, it was not possible to capture all of the surface irregularities by assuming specimen
symmetry. In other instances (Liang et al. 1997, Alshibli and Sture 1999, Sachan and Penumadu
2007), the entire specimen surface was not captured because only the zone of shear banding
within soil specimens was investigated and volumetric measurements of the entire specimen
were not obtained. Because all of these previous methods utilized external cameras, several
optical challenges were encountered, as described in detail in Bhandari et al. (2012). Zhang et al.
(2015) presented the first true photogrammetric local and total volume measurements of a
triaxial specimen. As stated in Zhang et al. (2015), the need for photogrammetry was based on
the significant limitations and unrealistic assumptions of other photograph-based measurement
methods (i.e., only local volume was obtained; accurate and precise control of relative camera
location and camera orientation were required). Although the Zhang et al. (2015) method
![Page 76: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/76.jpg)
60
overcame the limitations of many other photograph-based methods, the Zhang et al. (2015)
method still required computationally intensive ray tracing and least-squares optimization to
correct for the curvature of the cell wall, for the deformation of the cell under cell pressure, and
for light refraction at the confining fluid–cell wall and cell wall–atmosphere interfaces.
4.7. Internal Camera-Monitoring System
As described in Salazar and Coffman (2015), a small board camera with a pinhole
aperture (BCPA) device was developed to acquire photographs from within the triaxial cell. The
various challenges of utilizing internal photogrammetric instrumentation, namely, space
requirements, confining fluid, focal length, cell pressure, and specimen coverage, were overcome
by developing and utilizing the BCPAs. The optical, mechanical, and electrical components were
considered in the design of the BCPA device and combined BCPA system. Specifically, the
optical components were presented in Salazar and Coffman (2015), whereas the mechanical and
electrical components are presented herein. A schematic of the components of the combined
BCPA-monitoring system is presented in Figure 4.1.
![Page 77: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/77.jpg)
61
Figure 4.1. (a) Exploded view, and (b) elevation view of the internal components of the
combined BCPA monitoring system.
4.7.1. Mechanical Design
Multiple BCPA devices were employed to enable photographic coverage of the entire
surface of the specimen (during consolidation and shearing up to 15 % axial strain in triaxial
compression or triaxial extension). Given the vertical viewing angle of each of the individual
BCPAs (approximately 73º), the BCPA devices were stacked to allow for overlapping fields
of view along the length of the specimen in the axial direction. To minimize the required
number of BCPAs, a rotating platform was designed and fabricated to allow for several
overlapping images at each point around the specimen in the radial direction. Because of
the presence of two diametrically opposed vertical drain lines within the triaxial cell, which
enable drainage from the top of the specimen through the top platen, a full 360º revolution of
a single BCPA tower was not possible. Therefore, two diametrically opposed BCPA towers
4
7. Delrin® top guide
8. Tab for BCPA tower
9. Rotating Delrin®
bearing track
10. Delrin® base
11. Triaxial base
12. Cell wall
6
3
1. Drainage lines
2. Individual BCPA
device (10 total)
3. Acrylic cap (2 total)
4. BCPA tower
5. Soil specimen
6. Internal magnet
7
2
(a)
5
(b)
11
10
9 8
1
12
![Page 78: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/78.jpg)
62
were required, and the towers were rotated between the two drain lines. Each tower
rotated with 155º of rotation. Given the horizontal fields of view of the individual BCPA
devices and the required image overlap for photogrammetric processing, the towers were
rotated around the specimen and the towers were stopped at a desired degree interval to
acquire images.
To facilitate smooth and precise rotation of the two BCPA towers around the soil
specimen, the rotating platform utilized a stiff, low-friction, thermoplastic material
[polyoxymethylene (Delrin)], and an L-shaped slot design. A 25.4 mm x 25.4 mm x 12.7 mm
[1 in. x 1 in. x 0.5 in.] neodymium magnet (52 MGOe) located on the outside of the cell was
circumferentially rotated around the outside of the cell wall to pull a 6.35-mm [0.25-in.]
diameter neodymium magnet, which was mounted to the base of one of the BCPA towers,
causing the towers to rotate circumferentially around the vertical axis.
4.7.2. Electrical Design
Signals were passed into and out of the triaxial cell through the top cap of the cell by
utilizing a pinned throughput connector (as previously described in Salazar and Coffman
2014). Video, power, and ground wires were connected in such a way as to reduce the
required number of pin connections because of the large number of BCPA devices that were
employed (ten). The video and ground wires from each of the BCPA towers were connected
to common outputs, and the power wires from each BCPA were connected to a power control
switchboard (Figure 4.2), which allowed for only one BCPA to be powered at a given time
(thereby avoiding any video output feedback through the common grounding). Power was
individually supplied to each of the BCPA devices by increasing the amount of current that was
supplied to each transistor on the switchboard. Specifically, power was supplied to a given
![Page 79: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/79.jpg)
63
1. Capacitor (200 V, 100 μF)
2. Standard timer
3. Voltage regulator (1.2-37 V)
4. Power (BCPA devices, 3.7 V)
5. Transistors (222 A)
6. Sequencing counter (650 ns)
7. Capacitor (50 V, 10 μF)
8. Common ground
9. Power (switchboard, 6.0 V)
10. Variable resistor (timing, 10-20 sec.)
BCPA when the amount of current matched a given transistor. The user then acquired still
frames from the video feed when a desired BCPA device was powered. The power to the
switchboard was supplied via an AC to DC converter (6 V; maximum 1 A). The voltage
regulators on the switchboard conditioned the power to the requisite level (3.7 V) for the BCPA
devices. The video feed was collected using a desktop computer via a USB interface video
adapter (Sabrent USB-AVCPT). The still frames were captured and stored by utilizing video
software (Ulead VideoStudio). A wiring diagram of the entire data collection system is presented
in Figure 4.3.
Figure 4.2. Photograph of the power switch board for the BCPA devices (power supply and
timing sequence).
1
7
2 3 t
h
e
c
a
m
e
r
a
s
y
s
t
e
m
s
t
h
a
t
s
u
r
r
o
u
n
d
t
3
4
4
5 6
10
9 th
e
c
a
m
er
a
s
y
st
e
m
s
th
at
s
u
rr
o
u
n
d
th
e
tr
ia
xi
8 t
h
e
c
a
m
e
r
a
s
y
s
t
e
m
s
t
h
a
t
s
u
r
r
o
u
To BCPA devices
To throughput
connector
![Page 80: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/80.jpg)
64
Figure 4.3. Wiring diagram of the photogrammetric instrumentation.
4.8. Photogrammetric Methods and Results
The following process (Figure 4.4) was utilized to calibrate and validate the camera
monitoring system and to reconstruct a digital three-dimensional model of a brass test specimen
with nominal dimensions of 3.81 cm (1.5 in.) in diameter by 7.62 cm (3.0 in.) in length. (1) Each
BCPA device was calibrated by capturing a series of images of a printed PhotoModeler
Computer and image
acquisition program
Throughput pin connector
Universal serial bus
(USB) video adapter
Switchboard
Direct current
power supply to
switchboard and
BCPA devices
RCA video
connector
Video signal
BCPA devices/towers
Common ground splice
Power (BCPA devices)
Ground
(BCPA devices)
![Page 81: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/81.jpg)
65
calibration grid. The images of the calibration grid were subsequently analyzed using the
PhotoModeler Scanner software (PhotoModeler Scanner 2015) to obtain the necessary intrinsic
camera parameters (focal length, sensor format size, and principal point) of the BCPA. (2)
PhotoModeler coded targets [ringed automatically detected (RAD)] were adhered to the side of
the aforementioned brass specimen, and the specimen was placed on a flat surface where
additional RAD targets were adhered to the surface on which the specimen rested (96 targets
total). A previously calibrated digital single lens reflex (DSLR) camera (Canon 5D Mark II with
fixed 28mm Nikkor lens) with known properties (aperture controlled, f/13, ISO 1250, 28mm
fixed lens, and as calibrated using a PhotoModeler calibration grid) was used to capture images
of all sides of the specimen. Each RAD target was captured in a minimum of four images. (3)
The DSLR-acquired images were processed using the PhotoModeler software, and the center of
each target was precisely surveyed to within 0.2mm (0.0079 in.). (4) The same brass specimen,
with the same adhered targets, was then placed within the triaxial cell, and the target-covered
surface was captured using the BCPA devices that were mounted on the two towers. Five-degree
intervals were utilized to acquire a total of 320 images. As shown in Figure 4.5, adjacent images
were overlapped in both the axial and radial directions. (5) The common control points within
the DSLR and the BCPA images (see a12, b21, c32 in Figure 4.5) were used to
photogrammetrically derive the location and orientation of each of the individual BCPA devices,
at a given rotation interval, within the cell.
![Page 82: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/82.jpg)
66
Figure 4.4. (a) PhotoModeler camera calibration grid, (b) DSLR-acquired survey images
(control point identification), (c) BCPA-acquired calibration images (camera location and
orientation identification), and (d) BCPA-acquired images (point cloud identification).
Figure 4.5. Vertically and horizontally overlapping photographs captured with two
adjacent BCPA on a single tower at 15 degree rotation intervals.
a12
b21
a12 a12
b21
b21 b21
c32
c32
![Page 83: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/83.jpg)
67
To determine the practical range of rotation intervals (number of stations during rotation),
specific sets of the captured images were removed from the total 320 images (as captured from
the 5º rotation interval). Specifically, the locations and orientations of BCPA devices were
derived by using different intervals (45º, 30º, 15º, and 5º rotation intervals). Three-dimensional
recreations from the PhotoModeler software, of the calibration specimen surface (control points)
and BCPA device locations and orientations within the triaxial cell, are shown in Figure 4.6.
Photographic measurements obtained from the software, based on the DSLR survey, were
utilized to calculate a volume for the brass test specimen. A volume of 91.58 cm3 was obtained,
which corresponded to an estimated difference of 0.34 % when compared to the volume
calculations that were obtained from manual measurements (caliper and pi tape). Given
repeatable positioning of the BCPA towers at a desired rotation interval, the derived location and
orientation values for the individual BCPA devices were able to be used, in conjunction with
PhotoModeler software, to measure points on the surface at any axial strain level, for any soil
specimen that was tested within the triaxial cell. The measured points resulted in a point cloud
that was then used to identify the surfaces of the given specimen. The point cloud was then
exported from PhotoModeler Scanner and imported into Geomagic Design 3D software
(Geomagic Design X 2015) to obtain accurate threedimensional reconstructions of any specimen.
![Page 84: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/84.jpg)
68
Control point obtained from DSLR survey, used to identify the locations of the BCPA devices
Photograph from BCPA device location/rotation
Figure 4.6. Photogrammetric reconstruction of BCPA device locations within the triaxial
cell for (a) 45 degree, (b) 30 degree, (c) 15 degree, and (d) 5 degree rotation intervals.
(a) (b)
(c) (d) Location of drain line
Location of drain line
![Page 85: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/85.jpg)
69
4.9. Conclusions
Internal photogrammetric instrumentation was designed and implemented for triaxial
testing applications. The mechanical and electrical components of the BCPA instrumentation
were presented and the utilized photogrammetric techniques were discussed. The calibration and
validation processes for the system were also described. Based on preliminary findings, use of an
internal camera monitoring system is promising for future triaxial testing applications.
![Page 86: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/86.jpg)
70
4.10. References
Alshibli, K. A. and Al-Hamdan, M. Z., 2001, “Estimating Volume Change of Triaxial Soil
Specimens From Planar Images,” Comput.-Aided Civil Inf. Eng., Vol. 16, No. 6,
pp. 415–421.
Alshibli, K. A. and Sture, S., 1999, “Sand Shear Band Thickness Measurements by Digital
Imaging Techniques,” J. Comput. Civ. Eng., Vol. 13, No. 2, pp. 103–109.
Bésuelle, P. and Desrues, J., 2001, “An Internal Instrumentation for Axial and Radial Strain
Measurements in Triaxial Tests,” Geotech. Test. J., Vol. 24, No. 2, pp. 193–199.
Bhandari, A. R., Powrie,W., and Harkness, R. M., 2012, “A Digital Image-Based Deformation
Measurement System for Triaxial Tests,” Geotech. Test. J., Vol. 35, No. 2, pp. 209–226.
Bishop, A. W. and Donald, I. B., 1961, “The Experimental Study of Partly Saturated Soil in
Triaxial Apparatus,” Proceedings of the Fifth International Conference on Soil
Mechanics and Foundation Engineering, Vols. 1/3, Paris, July 17–22, Institution of Civil
Engineers, London, pp. 13–21.
Clayton, C. R. I., Khatrush, S. A., Bica, A. V. D., and Siddique, A., 1989, “The Use of Hall
Effect Semiconductors in Geotechnical Instrumentation,” Geotech. Test. J., Vol. 12,
No. 1, pp. 69–76.
Cuccovillo, T. and Coop, M. R., 1997, “The Measurement of Local Axial Strains in Triaxial
Tests Using LVDTs,” Géotechnique, Vol. 47, No. 1, pp. 167–171.
Gachet, P., Geiser, F., Laloui, L., and Vulliet, L., 2007, “Automated Digital Image Processing
for Volume Change Measurement in Triaxial Cells,” Geotech. Test. J., Vol. 30, No. 2,
pp. 98–103.
Geomagic Design X; Version 17 2015. “3D Computer-Aided Design (CAD) software,” 3D
Systems, Rock Hill, SC.
Gudehus, G. and Nübel, K., 2004, “Evolution of Shear Bands in Sand,” Géotechnique, Vol. 54,
No. 3, pp. 187–201.
Khan, M. H. and Hoag, D. L., 1979, “A Noncontacting Transducer for Measurement of Lateral
Strains,” Can. Geotech. J., Vol. 16, No. 2, pp. 409–411.
Leong, E. C., Agus, S. S., and Rahardjo, H., 2004, “Volume Change Measurement of Soil
Specimen in Triaxial Test,” Geotech. Test. J., Vol. 27, No. 1, pp. 47–56.
Liang, L., Saada, A., Figueroa, J. L., and Cope, C. T., 1997, “The Use of Digital Image
Processing in Monitoring Shear Band Development,” Geotech. Test. J., Vol. 20, No. 3,
pp. 324–339.
![Page 87: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/87.jpg)
71
Macari, E. J., Parker, J. K., and Costes, N. C., 1997, “Measurement of Volume Changes in
Triaxial Tests Using Digital Imaging Techniques,” Geotech. Test. J., Vol. 20, No. 1,
pp. 103–109.
Ng, C. W. W. and Chiu, C. F., 2001, “Behaviour of a Loosely Compacted Unsaturated Volcanic
Soil,” J. Geotech. Geoenviron. Eng., Vol. 127, No. 12, pp. 1027–1036.
Ng, C. W. W., Zhan, L. T., and Cui, Y. J., 2002, “A New Simple System for Measuring Volume
Changes in Unsaturated Soils,” Can. Geotech. J., Vol. 39, No. 3, pp. 757–764.
Nübel, K. and Weitbrecht, V., 2002, “Visualization of Localization in Grain Skeletons With
Particle Image Velocimetry,” J. Testing Eval., Vol. 30, No. 4, pp. 322–329.
Önal, O., Ören, A. H., Özden, G., and Kaya, A., 2008, “Determination of Cylindrical Soil
Specimen Dimensions by Imaging With Applications to Volume Change of Bentonite–
Sand Mixtures,” Geotech. Test. J., Vol. 31, No. 2, pp. 124–131.
Ören, A. H., Önal, O., Özden, G., and Kaya, A., 2006, “Nondestructive Evaluation of
Volumetric Shrinkage of Compacted Mixtures Using Digital Image Analysis,”
Eng. Geol., Vol. 85, No. 3, pp. 239–250.
PhotoModeler Scanner; Version 2015.0.0, 2015, “3D Measurements and Models Software,”
Eos Systems, New York.
Puppala, A. J., Katha, B., and Hoyos, L. R., 2004, “Volumetric Shrinkage Strain Measurements
in Expansive Soils Using Digital Imaging Technology,” Geotech. Test. J., Vol. 27,
No. 6, pp. 547–556.
Rechenmacher, A. L. and Finno, R. J., 2004, “Digital Image Correlation to Evaluate Shear
Banding in Dilative Sands,” Geotech. Test. J., Vol. 27, No. 1, pp. 13–22.
Rechenmacher, A. L. and Medina-Cetina, Z., 2007, “Calibration of Soil Constitutive Models
With Spatially Varying Parameters,” J. Geotech. Geoenviron. Eng., Vol. 133, No. 12,
pp. 1567–1576.
Sachan, A. and Penumadu, D., 2007, “Strain Localization in Solid Cylindrical Clay Specimens
Using Digital Image Analysis (DIA) Technique,” Soils Found., Vol. 47, No. 1,
pp. 67–78.
Salazar, S. E. and Coffman, R. A., 2014, “Design and Fabrication of End Platens for Acquisition
of Small-Strain Piezoelectric Measurements during Large-Strain Triaxial Extension and
Triaxial Compression Testing,” Geotech. Test. J., Vol. 37, No. 6, pp. 948–958.
doi:10.1520/GTJ20140057.
![Page 88: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/88.jpg)
72
Salazar, S. E. and Coffman, R. A., 2015, “Consideration of Internal Board Camera Optics for
Triaxial Testing Applications,” Geotech. Test. J., Vol. 38, No. 1, pp. 40–49.
doi:10.1520/GTJ20140163.
Scholey, G. K., Frost, J. D., Lo Presti, C. F., and Jamiolkowski, M., 1995, “A Review of
Instrumentation for Measuring Small Strains During Triaxial Testing of Soil Specimens,”
Geotech. Test. J., Vol. 18, No. 2, pp. 137–156.
Zhang, X., Li, L., Chen, G., and Lytton, R., 2015, “A Photogrammetry-Based Method to
Measure Total and Local Volume Changes of Unsaturated Soils During Triaxial
Testing,” Acta Geotech., Vol. 10, No. 1, pp. 55–82.
![Page 89: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/89.jpg)
73
CHAPTER 5: DISCUSSION OF "A PHOTOGRAMMETRY-BASED METHOD TO
MEASURE TOTAL AND LOCAL VOLUME CHANGES OF UNSATURATED SOILS
DURING TRIAXIAL TESTING" BY ZHANG ET AL., 2015
5.1. Chapter Overview
The Zhang et al. (2015) publication was discussed in this chapter. The Zhang et al. (2015)
manuscript contained a description of a photogrammetric method for determining total and local
volume changes of soil specimens during triaxial testing. Like the other camera-based triaxial
monitoring techniques found in the literature, Zhang et al. (2015) used cameras external to the
testing cell and therefore had to account for refraction effects and limited visibility of test
specimens through the cell wall. Despite this difference, the manuscript caught the attention of
the author, because of the similarities in the photogrammetric processing of results. This
manuscript echoes the Zhang et al. (2015) discussion of the limitations of non-photogrammetric
methods for determining triaxial specimen volume, describes the limitations of the Zhang et al.
(2015) work, and offers comparison with the methodology presented in Salazar and Coffman
(2015) and Salazar et al. (2015).
The full citation for this manuscript (Salazar and Coffman 2015) is included in Section
5.2. The discussion is presented in Section 5.3, which includes discourse on the apparent
improper triaxial testing techniques of the Zhang et al. (2015) publication (Section 5.3.1), a
discussion of the cell wall deformation corrections and least-square optimization (Section 5.3.2),
and photogrammetric methods (Section 5.3.3).
5.2. Discussion of "A Photogrammetry-based Method to Measure Total and Local Volume
Changes of Unsaturated Soils During Triaxial Testing" by Zhang et al., 2015
Reference
Salazar, S. E. and Coffman, R. A., “Discussion of ‘A Photogrammetry-based Method to Measure
Total and Local Volume Changes of Unsaturated Soils During Triaxial Testing’ by Zhang et al.”
Acta Geotechnica, Vol. 10, No. 5, pp. 693-696. doi:10.1007/s11440-015-0380-1.
![Page 90: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/90.jpg)
74
5.3. Discussion
Zhang et al. (2015), in the paper entitled ‘‘a photogrammetry-based method to measure
total and local volume changes of unsaturated soils during triaxial testing,’’ presented a method
for measuring the volume and strain of saturated and unsaturated soil specimens during triaxial
tests. Specifically, the presented method utilized a single, external digital camera, to capture
images of a triaxial testing apparatus and of the corresponding soil specimen within the cell.
Photogrammetric analyses were performed using commercially available photogrammetry
software. Utilizing the Zhang et al. (2015) modeling technique, ringed automatically detected
(RAD) coded targets were utilized to locate common points within each of the captured images,
and then, photogrammetry techniques were employed to assign physical, three-dimensional,
coordinates to each of the points.
A comprehensive and categorized review of triaxial volume measurement methods that
were found in the literature was presented in the Zhang et al. (2015) article. The methods that
have been previously utilized to measure the volume of saturated and unsaturated soil specimens,
during triaxial testing, were presented in a clear and concise manner by citing advantages and
disadvantages, accuracy, and costs associated with each method (Table 1 within Zhang et al.
(2015)). Furthermore, the current state of the art of photograph-based measurements was
explained. As stated in Zhang et al. (2015), the validity of other photograph-based volume
measurements [digital image analysis (DIA), digital image correlation (DIC), and particle image
velocimetry (PIV)] suffer from unrealistic, fundamental assumptions and limitations.
Specifically, the DIA methods require accurate and precise control of the relative camera
location and of the camera orientation relative to the triaxial cell soil specimen. The DIC and
PIV methods typically cannot provide total volume measurements; however, Bhandari et al.
![Page 91: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/91.jpg)
75
(2012) utilized the DIC method and overcame some of the difficulties associated with DIC
methods to provide viable results. However, the need for true photogrammetry to accurately
measure local and global volumes of saturated and unsaturated soil specimens was presented by
Zhang et al. (2015). Moreover, each of the steps of the photogrammetric process was thoroughly
explained. The methods presented in Zhang et al. (2015) are a valuable contribution to the
literature because the methods may be used to: measure total volume, measure local volume
changes (Figure 19 within Zhang et al. (2015)), and illustrate strain localization and shear
banding within soil specimens during triaxial testing. However, Zhang et al. (2015) do not
expound upon the limitations of the presented method. These limitations include the use of: (1)
improper triaxial testing procedures and (2) photogrammetric cell wall deformation
measurements combined with least-square optimization to obtain corrected ray path
measurements.
5.3.1. Improper Triaxial Testing Techniques
An internal load cell should be utilized in triaxial testing to prevent the need for piston
uplift and piston friction corrections. Silicone oil, instead of water, is commonly used within the
cell, as the cell fluid, when an internal load cell is utilized. The index of refraction for silicone oil
differs from the index of refraction for water, which may affect the observed results in a similar
manner to that shown for difference in the indices of refraction for air and water that is presented
in Figure 1 of the Zhang et al. (2015) article. Furthermore, because of the use of an external
camera, the cleanliness of the acrylic cell wall (which can be compromised within a soil
mechanics laboratory) may also affect the results.
Although Zhang et al. (2015) utilized an impressive back pressure saturation technique in
which the sand was infused with CO2 and then subjected to 400 kPa of back pressure to obtain a
![Page 92: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/92.jpg)
76
B value of 0.98, the back pressure saturation technique was for naught because the cell pressure
was reduced from 435 to 100 kPa and the back pressure was reduced from 100 to 0 kPa prior to
shearing. The reduction in cell pressure and back pressure most likely resulted in effervescence
of the pore fluid and therefore desaturation of the sand specimen to an unknown state of suction
within the soil specimen. Thereby, even though higher cell pressures (600 kPa) were utilized to
determine the accuracy of the method, these pressures were not utilized for the triaxial testing of
the soil specimens. This brings into question whether the Zhang et al. (2015) method will work if
the cell is pressurized under high cell pressures that are commonly required to back pressure
saturate and overconsolidate clay specimens (1035 kPa).
5.3.2. Utilization of Cell Wall Deformation Combined with Least-Square Optimization
The Zhang et al. (2015) method utilized images that were captured from outside of the
cell. Therefore, ray tracing was required to correct for the refraction effects of the light (1) at the
confining fluid–cell wall interface and (2) at the cell wall–air interface. Additionally, the method
considered deflection of the cell wall during pressurized tests by including RAD-coded targets
that were adhered to the outer surface of the cell wall and to the load frame (considered fixed
control points for reference purposes). It was assumed that the cell wall deformed in a uniform,
radial pattern. Furthermore, deformation of the cell wall was not considered above 600 kPa, even
though typical triaxial tests may reach confining fluid pressures above 1000 kPa. It is suggested
that the correction for cell wall flexure be characterized for a full range of pressures that are
commonly achieved during a typical triaxial test (up to 1035 kPa for acrylic cell walls).
Furthermore, it may be a useful contribution to show a sensitivity study of the photogrammetric
method as applied to multiple optical media.
![Page 93: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/93.jpg)
77
In Section 2.1.1, Zhang et al. (2015) identified the problems associated with the
utilization of the measurement of the cell fluid for determining the volume change of a triaxial
specimen. However, Zhang et al. (2015) utilized a similar method, albeit in the form of a
photogrammetric correction instead of a calibration procedure, to account for the change in the
ray path that is associated with the cell wall deformation during pressurization/depressurization
of the cell wall deformation under a constant applied pressure (creep). Moreover, the camera that
was utilized was a lensed camera. Before analysis of captured images, photogrammetric software
was used by Zhang et al. (2015) to correct for lens distortions (thereby modeling the lensed
camera as a pinhole camera).
Recently, as presented in Salazar and Coffman (2015a, 2015b), a triaxial insert board
camera pinhole aperture (BCPA) camera system was developed by researchers at the University
of Arkansas. Although utilization of eight cameras was presented in the Salazar and Coffman
(2015a) article, the system was further modified to consist of ten small cameras (five cameras per
tower, with the two towers being diametrically opposed), as presented in Salazar and Coffman
(2015b). Two towers were required because the drainage lines for the top platen prevented a full
360º rotation of only one tower. Therefore, using the BPCA tower system, ten photographs were
acquired at a given position, while each of the towers completed a 155º rotation around the
specimen. The BCPA system was fully submerged in the silicone confining fluid, and the BCPA
was capable of being saturated and pressurized under pressures of up to 1035 kPa. Therefore, the
BCPA system enabled testing conditions that were similar to those that are commonly used
instead of requiring that the cell pressure and back pressure be reduced prior to shearing. The use
of the BPCA camera system: (1) reduces the complicated geometry of the camera location and
orientation by utilizing photogrammetry to derive the exact camera location and orientation, (2)
![Page 94: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/94.jpg)
78
does not require determination of the best fit of the shape and location of the acrylic cell, (3)
does not require ray tracing or Snell’s law because the photographs are acquired from within the
cell fluid, and (4) provides an alternative to externally acquired photogrammetric methods.
5.3.3. Testing Procedures and Photogrammetric Methods
Zhang et al. (2015) performed isotropic compression tests for a stainless steel specimen
to measure the accuracy for the photogrammetry method. Tests were performed on a 5.08 cm
diameter by 10.16 cm tall stainless steel specimen that was tested within a 10.16 cm diameter by
20.32 cm tall acrylic cell with a 0.61 cm thick cell wall that had a refractive index of 1.491.
Approximately 50 photographs were acquired for each testing condition (0 kPa in air without cell
wall; 0, 200, 400, 600 kPa in water), by taking at least five photographs from different
orientations for each area/point of interest to ensure ‘‘sufficient overlap between adjacent
pictures.’’ Targets numbering 16, 218, and 336 were utilized to identify the load frame, acrylic
cell, and stainless steel specimen, respectively.
Drained triaxial tests were also performed on a saturated sand specimen to determine the
total and local volume measurements of the specimen during the triaxial tests. Unlike the triaxial
cell that was utilized for the isotropic compression tests on the stainless steel cylinder, a larger
triaxial cell with larger specimens and fewer targets was used for the triaxial tests conducted on
the saturated sand specimens. Specifically, a 7.1 cm by 13.7 cm specimen was tested within a
16.51 cm diameter by 30.48 cm tall triaxial cell that had a cell wall thickness of 0.97 cm and a
refractive index of 1.491. Also, 174 and 176 targets were utilized to identify the acrylic cell and
sand specimen, respectively, within the 25 photographs that were acquired for each testing
condition (at every 2–3 mm of vertical displacement…until a total displacement of 15 mm is
reached). The reason for using 336 targets adhered to the specimen was never explained, nor
![Page 95: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/95.jpg)
79
justified. Furthermore, it was not clear how point capture redundancy eliminates any assumptions
regarding the specimen deformations. This statement requires further explanation. A parametric
study of the accuracy achieved, for different numbers of points utilized, would be a useful
contribution to the literature and would help other researchers make decisions on the level of
refinement required to achieve a desired level of accuracy. Also, a parametric study would aid in
determining the minimum number of targets required to measure the total and local volumes of
specimens. Although the results presented by Zhang et al. (2015) would indicate that the method
is capable of achieving high accuracy (<0.25 % error), the method is computationally intensive,
due to the effects of refraction and cell wall deformation.
At the University of Arkansas, researchers performed tests on a brass specimen with
nominal dimensions of 3.8 cm in diameter and 7.6 cm in height that contained 273 targets;
thereby, the University of Arkansas specimen was smaller than the specimen used by Zhang et
al. (2015) and contained more targets on the surface of the soil specimen. Each of the drainage
lines prevented direct viewing of 25º of the specimen; however, the photos collected near the
drainage lines allowed for points within these locations to be viewed in at least six photos,
instead of the customary ten photos that were used for the other points. In addition to the tests on
a brass specimen, at various times during triaxial tests on soil specimens [prior to confinement,
during back pressure saturation, prior to consolidation, during consolidation, after consolidation,
and during shearing (1, 3, 6, 9, 12, and 15 % axial strain)], photographs were obtained by
rotating the towers at a desired increment ranging from a minimum of 45º increments to a
maximum of 5º increments. Specifically, ranging from a minimum of 40 pictures per observation
(10 pictures per increment, and 4 increments) to a maximum of 320 pictures per observation (10
pictures per increment, and 32 increments), respectively.
![Page 96: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/96.jpg)
80
The same procedures that were utilized by Zhang et al. (2015) for the image analysis in
air are common accuracy prediction methods, and these procedures were also utilized by Salazar
and Coffman (2015b). However, the Salazar and Coffman (2015b) method offered a viable
alternative to the Zhang et al. (2015) method, by placing photogrammetric equipment within the
triaxial cell instead of outside of the triaxial cell. The lensless camera equipment that was
devised by Salazar and Coffman (2015a) allowed for direct, unobstructed observation of a
specimen during testing. Therefore, computationally intensive corrections to account for the
effects of refraction through multiple types of media and deformation of the cell wall were
eliminated. In a similar fashion to the Zhang et al. (2015) method, the Salazar and Coffman
(2015b) method utilized the same principles of photogrammetry (using PhotoModeler Scanner
software (Eos Systems, Inc. (2015)) and RAD-coded targets to create a point cloud of the
specimen surface. In the Salazar and Coffman (2015b) method, the point cloud was then meshed
to obtain an accurate, three-dimensional reconstruction of the specimen, whereupon the mesh
was imported into Geomagic Design software (3D Systems, Inc., 2015) to calculate the volume
of the specimen. The volumetric strain was then obtained by subtracting the volume of the
specimen at various times from the initial volume and then dividing by the initial volume.
Whereas Zhang et al. (2015) utilized a proprietary program (PhotoSoilVolume) to handle the
cumbersome optical correction process through to volume calculation, the Salazar and Coffman
(2015b) method required only commercially available software to calculate total specimen
volumes. As described in Salazar and Coffman (2015a), the optical components of internal
photogrammetric instrumentation were designed, fabricated, and tested for triaxial testing
applications to overcome the aforementioned drawbacks of external methods. The cameras were
![Page 97: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/97.jpg)
81
also designed to overcome the challenges of internal instrumentation (space, confining fluid,
focal length, cell pressure, and specimen coverage).
In summary, the contribution produced by Zhang et al. (2015) is significant and provides
an advancement of the state of knowledge of external photograph-based measurements of triaxial
specimens. However, the use of internal cameras like those presented in Salazar and Coffman
(2015a, 2015b), instead of an external camera, is suggested to further advance the
photogrammetric technique. Specifically, the use of internal cameras will facilitate a reduction in
computational demands by (1) eliminating the need for a cell deformation/position correction
and by (2) eliminating the need for ray tracing and least-square optimization. Although the
Salazar and Coffman (2015b) method is less computationally costly than the method presented
by Zhang et al. (2015), the cost and reproducibility of required photogrammetric equipment are
approximately equivalent. Therefore, it appears that the Salazar and Coffman (2015a) method is
a viable alternative method for acquiring total and local volume changes of soil specimens during
triaxial testing.
![Page 98: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/98.jpg)
82
5.4. References
3D Systems, Inc., 2015, Geomagic Design X (Version 17). 3D Computer-Aided Design
(CAD) Software. http://www.geomagic.com/en/products/design/overview
Bhandari, A. R., Powrie, W., and Harkness, R. M., 2012, “A Digital Image-Based Deformation
Measurement System for Triaxial Tests,” Geotech. Test. J., Vol. 35, No. 2, pp. 1-18.
Eos Systems, Inc., 2015, PhotoModeler Scanner (Version 2015.0.0). 3D Measurements and
Models Software. http://www.photomodeler.com/products/scanner/default.html
Salazar, S. E., Barnes, A., and Coffman, R. A., 2015b, “Development of an Internal Camera
Based Volume Determination System for Triaxial Testing,” Geotech. Test. J., Vol. 38,
No. 4, pp. 549-555. doi:10.1520/GTJ20140249.
Salazar, S. E. and Coffman, R. A., 2015a, “Consideration of Internal Board Camera Optics for
Triaxial Testing Applications,” Geotech. Test. J., Vol. 38, No. 1, pp. 40-49.
doi:10.1520/GTJ20140163.
Zhang, X., Li, L., Chen, G., and Lytton, R., 2015, “A Photogrammetry-Based Method to
Measure Total and Local Volume Changes of Unsaturated Soils During Triaxial
Testing,” Acta Geotech., Vol. 10, No. 1, pp. 55-82. doi:10.1007/s11440-014-0346-8.
![Page 99: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/99.jpg)
83
CHAPTER 6: CLOSURE TO "DISCUSSION OF 'DEVELOPMENT OF AN
INTERNAL CAMERA-BASED VOLUME DETERMINATION SYSTEM FOR
TRIAXIAL TESTING' BY S. E. SALAZAR, A. BARNES, AND R. A. COFFMAN"
BY MEHDIZADEH ET AL., 2016
6.1. Chapter Overview
A closure to the Mehdizadeh et al. (2016) discussion paper is provided in this chapter.
The closure addresses queries that were raised by the discussion paper and clarifies the apparent
ambiguities of the previously described Salazar et al. (2015) technique. The manuscript also
provides additional testing data that shed light on the effect of pausing a triaxial test for short
periods of time (at desired axial strain intervals) to capture photographs of the surface of the
specimen. The full citation for this manuscript (Salazar et al. 2017) is included in Section 6.2 and
the body of the closure is presented in Sections 6.3 and 6.4.
6.2. Closure to “Discussion of ‘Development of an Internal Camera-Based Volume
Determination System for Triaxial Testing’ by S.E. Salazar, A. Barnes, and R.A. Coffman”
by Mehdizadeh et al., 2016
Reference
Salazar, S. E., Barnes, A., and Coffman, R. A., “Closure to “Discussion of ‘Development of an
Internal Camera-Based Volume Determination System for Triaxial Testing’ by S. E. Salazar, A.
Barnes, and R. A. Coffman” by A. Mehdizadeh, M. M. Disfani, R. Evans, A. Arulrajah, and D. E.
L. Ong,” Geotechnical Testing Journal, Vol. 40, No. 1, 2017, pp. 47-51.
doi:10.1520/GTJ20160154.
6.3. Abstract
The discussion paper, written by Mehdizadeh et al., provided discourse on a
technical note entitled “Development of an Internal Camera-Based Volume Determination
System for Triaxial Testing,” which presented a novel technique of photographically
monitoring soil specimens during triaxial tests by utilizing small board cameras internal to
the triaxial cell. Here, a closure to the discussion paper was provided; queries raised by the
![Page 100: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/100.jpg)
84
discussers were addressed, ambiguities were clarified, and additional testing data to support
the closure were provided.
Keywords: triaxial testing, photogrammetry, volume measurement, local deformation
6.4. Introduction
The authors appreciate Mehdizadeh et al. (2016), herein after referred to as the
discussers, for their interest in the technical note entitled “Development of an Internal
Camera-Based Volume Determination System for Triaxial Testing.” The authors believe
that the discussers (1) primarily wished to highlight the ambiguities contained within the
technical note, while also pointing to the work of Uchaipichat et al. (2011), and (2) offered a
simple solution to deal with optical distortion corrections. In this closure, the authors
addressed the comments put forth by the discussers, in the order in which they appeared in
the discussion paper, to clarify the ambiguities within the original technical note.
The discussers mentioned that Uchaipichat et al. (2011) monitored the volumetric
strain of a soil specimen during a triaxial compression test with the aid of only two cameras
(external to the confining cell). However, the authors agree with Uchaipichat et al. (2011)
that the digital image analysis method that was employed by Uchaipichat et al. (2011)
cannot be used to reliably determine the specimen volume in the case of nonuniform
deformation. The volume cannot be accurately measured using this technique because the
camera setup does not capture localized strain (such as observed in shear banding) and relies
heavily on averaging around the circumference of the specimen. While the Uchaipichat et al.
(2011) method may be accurate enough for some applications (such as determining the total
volumetric strain or approximating the radial deformation at mid-height), the technique may
not be suitable for applications where it is desired to determine absolute sample volume or
![Page 101: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/101.jpg)
85
to characterize the development of local deformations (by tracking individual targets on the
specimen surface). Although Uchaipichat et al. (2011) showed that the averaging technique
correlated well with the burette measurements for the test presented in the manuscript, the
use of only one frontal view image and one side view image cannot do justice in every case,
especially in the case of non-uniform deformation.
The discussers’ demonstration that the need to account for the “distortion of light,
cell curvature, and camera lens” could be eliminated altogether by basing all relative
measurements on a point of reference is intriguing. The authors commend Uchaipichat et al.
(2011) and the discussers for avoiding cumbersome corrections for optical distortions;
however, this simple calibration procedure can only be used to determine changes in total
specimen volume relative to an a priori volume. It is not clear how the initial volume of the
specimen would be obtained with sufficient accuracy to allow for any absolute
measurements before, during, or after the test. Any initial volume measurements, however
meticulous, would be for naught once the specimen was placed inside of the cell and set up
for testing (i.e., piston contact with specimen, filling of the cell with confining fluid).
Furthermore, it is not clear how the discussers suggest to correct for cell wall deformation
during tests when the confining pressure in the cell changes throughout the stages of testing
(during back pressure saturation, consolidation, and shearing). As listed in Table 1 in
Salazar and Coffman (2014), there are four stress paths where the cell pressure changes
during the shearing stage. Moreover, Zhang et al. (2015a) and Salazar and Coffman (2015b)
documented that the amount of cell wall deformation might be a significant factor when
tracing rays through the media surrounding the specimen (confining fluid, cell wall, air).
![Page 102: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/102.jpg)
86
The discussers noted correctly that due to the limited field of view for each of the
board camera devices that were presented in Salazar et al. (2015), the size of the observed
specimens was limited to nominal dimensions of 38.1-mm in diameter and 76.2-mm in
height. The triaxial stress path testing performed at the University of Arkansas is primarily
for stress-strain behavior to derive parameters required for constitutive modeling. The
specimens that are typically tested are primarily soft soils that have been (1) trimmed from a
Shelby tube sample, or (2) reconstituted in one of the 38.1-mm diameter, static weight slurry
consolidometers (in a similar fashion to Zhao and Coffman (2016) and Zhao et al. (2016))
prior to K0 reconsolidation within the triaxial cell. Furthermore, this smaller specimen size is
favored because it allows for reduced consolidation time over larger specimen sizes. It
should alleviate the discussers’ concern to know that the internal photogrammetry system
could easily be adapted to larger cells and testing of larger specimens. In fact, a larger cell
would allow for more space between the specimen and camera tower, increasing the vertical
and horizontal area viewed by each camera. This could also serve to reduce the number of
internal board cameras required, thereby reducing the number of images captured and the
amount of photogrammetric processing required.
The system presented in Salazar and Coffman (2015a, 2015b) and Salazar et al.
(2015) is neither overly complicated nor costly (especially as compared to a DSLR camera
and lens). However, it does require knowledge of photogrammetry for implementation. The
term “photogrammetry” was misused throughout the discussion by Mehdizadeh et al.
(2016); the term “photogrammetry” was interchanged with other photographic methods
(including digital image analysis and digital image correlation). Salazar and Coffman
(2015a, 2015b), Salazar et al. (2015), as well as Zhang et al. (2015a, 2015b), have proposed
![Page 103: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/103.jpg)
87
that the application of photogrammetry (or more accurately, stereophotogrammetry) is more
robust than other digital imaging techniques for determining the total- and local- radial,
axial, and volumetric strains. Specifically, photogrammetry overcomes the weaknesses of
other photographic methods while becoming increasingly easier to apply with the increasing
availability of various commercial and open-source photogrammetry processing software
packages. In addition, other photogrammetric methods, including open-source structure
from motion (SfM) and multi-view stereo (MVS) applications, using more recently
developed computer vision techniques, merit further investigation. Unlike the
stereophotogrammetry technique that was discussed in Salazar et al. (2015,2016), no prior
camera calibrations would be required using the MVS technique. Therefore, overall
processing times (even for the relatively large number of images captured with the internal
camera based system) would be dramatically reduced, and the reconstructed 3D model
would contain more surface detail (Furukawa and Hernández 2013).
The discussers also mentioned that the use of silicone oil for the confining fluid, in
place of water, would necessitate the use of specialized flow pumps that are not commonly
available. In fact, the adaptation of silicone oil as confining fluid has become commonplace
in research laboratories that utilize internal load cells, and there are several triaxial apparatus
manufacturers that sell flow pumps for this very purpose. As an aside, the internal load cells
are necessary in advanced triaxial stress path testing programs to account for the effects of
piston uplift and piston friction (Race and Coffman 2011, Salazar and Coffman 2015b). The
authors agree with the discussers that the choice of confining fluid has no particular impact
on external photography of the triaxial specimen (as long as the fluid is not opaque and the
index of refraction is taken into account). It was previously highlighted that the adaptation
![Page 104: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/104.jpg)
88
of open-body internal camera equipment was made possible by the already in-place practice
of using inert silicone oil as a confining fluid. The oil allows the camera devices and
associated electronics to be fully saturated and fully submerged in the confining fluid
without becoming damaged. Furthermore, the open-body style of the board cameras has
been proven to eliminate any pressure differentials by allowing the fluid to fill the lensless
cameras, including behind the pinhole aperture (as explained in Salazar and Coffman
2015a). This open body is a key design element of the camera devices, as it would be
impractical (and perhaps even impossible) to design compact, waterproof camera housings
capable of withstanding the pressures that are present within the confining cell during a
typical triaxial test. Furthermore, due to the lack of difference in indices of refraction
between the lens and the fluid, the authors’ discussion of confining fluid, as related to
refraction, was only provided to highlight the fact that the use of a lens submerged within
the fluid does not function as intended. Specifically, the index of refraction of a typical lens
is around 1.5 (depending on the material), which is too close to the index of refraction of
water (around 1.33) or silicone oil (1.397, as tested). The refraction of light through a lens in
air requires a larger difference (i.e. the index of refraction of air is approximately 1.0).
Therefore, the use of a lensless design was necessary to function within the cell; the
elimination of typical distortions associated with lenses was a fortuitous side-effect,
resulting in less corrections during photogrammetric calibration and processing. The use of a
lensless camera design is not without problems, due to the necessity of having good lighting
to allow the camera sensors to collect data and to reduce vignetting at the edges of the
image. However, proper lighting is still a key factor in collecting quality data when using
external DSLR photography.
![Page 105: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/105.jpg)
89
The authors commend the discussers for their approach to demonstrating a simple
and practical method of monitoring a triaxial specimen from the exterior of the cell, by
following the Uchaipichat et al. (2011) method. Particularly, the authors appreciate the
discussers’ use of stainless steel calibration specimens with known dimensions for
reference, as this practice is currently missing from the available body of literature for
photographic observation of triaxial specimens. In a similar fashion, the authors have
performed multiple comparison tests using brass and acrylic specimens with known
dimensions to validate the internal photogrammetry technique. The results of these tests
were documented in Salazar et al. (2016). Furthermore, the authors have continued on to
demonstrate the capabilities of the internal photogrammetry approach by monitoring soil
specimens during triaxial compression and extension tests (Salazar et al. 2016). Because no
universal comparison exists to assess the error for new volume measurement techniques
(and often there is no comparison to an external reference at all), the authors have attempted
to evaluate the difference in measured volume from a reference measurement. In Table 1 of
Salazar et al. (2016), a series of volume measurements for an acrylic specimen was
presented. The volume of the acrylic specimen was determined using a water displacement
technique (by adapting the Proctor mold volume determination method from ASTM D698-
12e2), manual measurements using a pi tape and calipers, 3D scan measurements, external
DSLR photogrammetry (camera located outside of cell, in air), and internal photogrammetry
(cameras located in cell filled with silicone oil confining fluid). The measured volume of the
acrylic specimen, determined using the internal photogrammetry approach, fell within 0.13
% difference from the reference volume. Although the reference measurement technique for
this example was the water displacement technique (because it was based on an ASTM
![Page 106: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/106.jpg)
90
standard), the measured volumes of the acrylic specimen fell within 0.5 % of the reference
for all five volume determination techniques.
The discussers commented that the Salazar et al. (2015) technique required rotation
of the camera tower platform around the specimen to capture images of the entire surface.
To ensure that there were no changes to the specimen during the image-capturing process,
the test was paused at the desired intervals. The authors concede that the acquisition of the
necessary images may be time-consuming; however, the short pauses during a test did not
cause problems in the stress path for the test (Figures 6.1 and 6.2) and excess pore water
pressure returned to the prescribed behavior after each pause during the test (Figure 6.3).
Furthermore, an advantage of the technique is that the camera towers may be rotated to
occupy any desired rotation interval around the specimen. Therefore, the acquisition time
and subsequent processing time may be significantly reduced with a reduced number of
images (i.e., greater angles between the camera intervals).
![Page 107: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/107.jpg)
91
Figure 6.1. Comparison of deviatoric stress as a function of mean stress for a reduced
triaxial extension test with pauses and without pauses for capturing photographs.
Figure 6.2. Comparison of deviatoric stress as a function of axial strain for a reduced
triaxial extension test with pauses and without pauses for capturing photographs.
-200
-150
-100
-50
0
50
100
150
200
0 50 100 150 200 250 300 350 400
Dev
iato
ric
Str
ess,
q =
σ1
-σ
3 ,
[kP
a]
Mean Stress, p' = (σ1' + 2σ3')/3, [kPa]
RTE Test with Pauses for Photographs
RTE Test without Pauses for Photographs
-200
-150
-100
-50
0
50
100
150
200
-12-10-8-6-4-20
Dev
iato
ric
Str
ess,
q =
σ1
-σ
3 ,
[kP
a]
Axial Strain, εa , [%]
RTE Test with Pauses for Photographs
RTE Test without Pauses for Photographs
Specimen Parameters:
wc = 50%, OCR = 1,
310 kPa K0 Consolidation
Specimen Parameters:
wc = 50%, OCR = 1,
310 kPa K0 Consolidation
![Page 108: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/108.jpg)
92
Figure 6.3. Comparison of excess pore water pressure development as a function of axial
strain for a reduced triaxial extension test with pauses and without pauses for capturing
photographs.
It was mentioned that during the photographic reconstruction of the stainless steel
weights, the discussers experienced some difficulty establishing the specimen edges within
the cell. The advantage of the photogrammetry techniques presented in Zhang et al. (2015a)
and Salazar et al. (2015) is that the long edges (profile) of the specimen need never be
established, because photogrammetry allows for the triangulation of individual points on the
surface of the specimen in 3D space. However, similar to the discussers’ experience, the
authors have also had some difficulty in establishing reliable points on the ends (top and
bottom) of the specimen. This difficulty was overcome by manually picking common
marker points, within each photo, along the top and bottom edges of the specimen within the
photogrammetry software.
-75
-50
-25
0
25
-12-10-8-6-4-20
Ex
cess
Po
re P
ress
ure
, Δ
u,
[kP
a]
Axial Strain, εa , [%]
RTE Test with Pauses for Photographs
RTE Test without Pauses for Photographs
Specimen Parameters:
wc = 50%, OCR = 1,
310 kPa K0 Consolidation
![Page 109: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/109.jpg)
93
As mentioned in Salazar and Coffman (2015b), the authors share the discussers’
concerns that cell wall cleanliness and obstruction due to triaxial cell apparatus (frame rods,
cell wall confining rings) are factors not to be overlooked when photographically
monitoring specimens from the exterior of the cell. However, the discussers successfully
demonstrated that the volumetric strain of a triaxial sample could be determined to a
reasonable level of accuracy without having to take into account cumbersome corrections:
(1) for optical distortions due to light refraction at the confining fluid-cell wall and cell wall-
air interfaces, and (2) due to cell wall curvature. The authors suggest that for future
applications, where only relative measurements are of interest to the discussers, more
camera positions should be utilized within the Uchaipichat et al. (2011) methodology to
capture more images of the specimen to increase the accuracy of non-uniform volumetric
strain measurements.
![Page 110: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/110.jpg)
94
6.5. References
ASTM D698-12e2, 2012, "Standard Test Methods for Laboratory Compaction
Characteristics of Soil Using Standard Effort (12 400 ft-lbf/ft3 (600 kN-m/m3))",
ASTM International, West Conshohocken, PA, www.astm.org.
doi:10.1520/D0698-12E02.
Furukawa, Y. and Hernández, C., 2013, “Multi-View Stereo: A Tutorial,” Found. Trends
Comput. Graphics Vis., Vol. 9, Nos. 1–2, pp. 1–148.
Mehdizadeh, A., Disfani, M. M., Evans, R., Arulrajah, A., and Ong, D. E. L., 2016,
“Discussion of ‘Development of an Internal Camera-Based Volume Determination
System for Triaxial Testing’ by S. E. Salazar, A. Barnes and R. A. Coffman,”
Geotech. Test. J., Vol. 38, No. 4, 2015, pp. 549–555.
Race, M. L. and Coffman, R. A., 2011, “Effects of Piston Uplift, Piston Friction, and
Machine Deflection in Reduced Triaxial Extension Testing,” ASCE Geotechnical
Special Publication No. 211, Proceedings GeoFrontiers 2011: Advances in
Geotechnical Engineering, ASCE, Reston, VA, pp. 2649–2658.
Salazar, S. E., Barnes, A., and Coffman, R. A., 2015, “Development of an Internal Camera-
Based Volume Determination System for Triaxial Testing,” Geotech. Test. J.,
Vol. 38, No. 4, pp. 549–555.
Salazar, S. E. and Coffman, R. A., 2014, “Design and Fabrication of End Platens for
Acquisition of Small-Strain Piezoelectric Measurements During Large-Strain
Triaxial Extension and Triaxial Compression Testing,” Geotech. Test. J., Vol. 37,
No. 6, pp. 948–958. doi:10.1520/GTJ20140057.
Salazar, S. E. and Coffman, R. A., 2015a, “Consideration of Internal Board Camera Optics
for Triaxial Testing Applications,” Geotech. Test. J., Vol. 38, No. 1, pp. 40–49.
doi:10.1520/GTJ20140163
Salazar, S. E. and Coffman, R. A., 2015b, “Discussion of ‘A Photogrammetry-Based
Method to Measure Total and Local Volume Changes of Unsaturated Soils During
Triaxial Testing’ by Zhang et al.,” Acta Geotech., Vol. 10, No. 5, pp. 693–696.
Salazar, S. E., Miramontes, L. D., Barnes, A., Bernhardt, M. L., and Coffman, R. A., 2016,
“An Internal Close-Range Photogrammetry Approach to Volume Determination for
Triaxial Testing,” Acta Geotech. (under review).
Uchaipichat, A., Khalili, N., and Zargarbashi, S., 2011, “A Temperature Controlled Triaxial
Apparatus for Testing Unsaturated Soils,” Geotech. Test. J., Vol. 34, No. 5,
pp. 424–432.
![Page 111: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/111.jpg)
95
Zhang, X., Li, L., Chen, G., and Lytton, R., 2015a, “A Photogrammetry-Based Method to
Measure Total and Local Volume Changes of Unsaturated Soils During Triaxial
Testing,” Acta Geotech., Vol. 10, No. 1, pp. 55–82.
Zhang, X., Li, L., Chen, G., and Lytton, R., 2015b, “Reply to ‘Discussion of “A
Photogrammetry-Based Method to Measure Total and Local Volume Changes of
Unsaturated Soils During Triaxial Testing” by Zhang et al.’ by Salazar and
Coffman,” Acta Geotech., Vol. 10, No. 5, pp. 697–702.
Zhao, Y. and Coffman, R. A., 2016, “Back-Pressure Saturated Constant-Rate-of-Strain
Consolidation Device With Bender Elements: Verification of System Compliance,”
J. Test. Eval., Vol. 44, No. 6, pp. 2375–2386.
Zhao, Y., Mahmood, N. S., and Coffman, R. A., 2016, “Small-Strain and Large-Strain
Modulus Measurements With a Consolidation Device,” J. Test. Eval. (under review).
![Page 112: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/112.jpg)
96
CHAPTER 7: VERIFICATION OF AN INTERNAL CLOSE-RANGE
PHOTOGRAMMETRY APPROACH FOR VOLUME DETERMINATION
DURING TRIAXIAL TESTING
7.1. Chapter Overview
A comprehensive description of the internal camera-based photogrammetry technique for
triaxial testing, that was developed at the University of Arkansas (Salazar and Coffman 2015a,
2015b, 2016; Salazar et al. 2015, 2017), is presented in this chapter. A series of sensitivity
studies were performed to evaluate the accuracy, precision, and feasibility of the technique.
Furthermore, the technique was implemented for two undrained triaxial tests on soil specimens.
Specifically, one conventional triaxial compression (CTC) and one reduced triaxial extension
(RTE) test were performed on slurry-consolidated, reconstituted kaolinite soil specimens. Results
from these tests, discussion of the limitations of the technique, and anticipated improvements to
the technique are presented.
The limitations of the included manuscript (Salazar et al. 2017) are discussed in
Section 7.2. The full citation for this document is included in Section 7.3. The motivation and
background for the manuscript are described in Sections 7.4 and 7.5. Section 7.6 presents the
evaluation of the internal photogrammetry technique. Specifically, the calibration of board
cameras, derivation of camera stations, determination of photograph intervals, capture of
photographs within confining fluid, photogrammetric reconstruction, and volume determination
are presented in Sections 7.6.1 through 7.6.6. Contained within Section 7.6.7 is a description of
the method used to determine the accuracy of the internal photogrammetry technique, which
included comparison of results from five different volume determination methods. Limitations
and sources of error are discussed in Section 7.6.8. The methods that were used for
![Page 113: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/113.jpg)
97
implementation of the technique during triaxial compression and triaxial extension tests are
detailed in Section 7.7. Results from the sensitivity studies and from the triaxial tests are
discussed in Section 7.8. Finally, conclusions are presented in Section 7.9.
7.2. Limitations of the Described Study
Although the manuscript contained within this chapter clarifies and further expounds on
the technique that was introduced in preceding publications, there are limitations to the work.
The various sensitivity studies that were completed as part of the evaluation of the technique
involved primarily analog (dummy) specimens. The manuscript also described the
implementation of the technique for two triaxial tests; however, these tests were undrained. It
was therefore not possible to compare changes in specimen volume, as determined using the
photogrammetry technique, with changes in pore fluid volume, as measured by the pore fluid
pump. Furthermore, the work only describes one triaxial compression and one triaxial extension
test. No repeat tests were performed to determine the precision of the technique for a given
triaxial test.
7.3. Verification of an Internal Close-Range Photogrammetry Approach for Volume
Determination During Triaxial Testing
Reference
Salazar, S. E., Miramontes, L. D., Bernhardt, M. L., and Coffman, R. A., “Verification of an
Internal Close-Range Photogrammetry Approach for Volume Determination During Triaxial
Testing,” Geotechnical Testing Journal. Submitted for Review. Manuscript Number: GTJ-2017-
0125.R2.
7.4. Abstract
Accurate strain and volume measurements are critical to phase relationships and strength
determination for saturated and unsaturated soils. In recent years, laboratory-based photographic
techniques of monitoring soil specimens have become more common. These techniques have
been used to reconstruct 3D models and to determine strain and volumetric changes of triaxial
![Page 114: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/114.jpg)
98
specimens. A new technique that used digital photographs of the soil specimen, captured from
within a triaxial testing cell, was utilized. Photographs were processed using photogrammetric
software to reconstruct 3D models of the soil specimens at various times during the triaxial test.
By placing camera equipment within the cell, the technique eliminated the need to account for
optical distortions due to 1) refraction at the confining fluid-cell wall-atmosphere interface,
2) the curvature of the cylindrical cell wall, and 3) the pressure-induced deformation of the cell
wall.
Previously unreported results from sensitivity studies and accuracy measurements for the
internal photogrammetry approach are documented herein. Furthermore, through undrained
triaxial compression and extension tests, the viability of determining total and local strains,
volume changes, and total volume at any stage of triaxial testing was demonstrated. By
comparison with other volume-determination methods that are presented herein, including DSLR
camera photogrammetry, 3D scanning, manual measurements, and water displacement
techniques, a relative error of the internal photogrammetry technique of 0.13 percent was
assessed.
Keywords: triaxial testing, photogrammetry, volume measurements, local deformation
7.5. Introduction and Background
Researchers have employed various methods (both photograph- and non-photograph-
based) for monitoring soil specimens during triaxial tests. Specifically, these measurements have
enabled one or more of the following: 1) axial and radial dimensions and deformations with time,
2) local and/or total volume measurements, 3) volumetric strain calculations, and 4) shear band
characterization. Examples include double-wall cell systems, differential pressure transducers,
measurements of air and water volume changes (Bishop and Donald 1961, Ng et al. 2002, Leong
![Page 115: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/115.jpg)
99
et al. 2004), displacement sensors (Scholey et al. 1995, Bésuelle and Desrues 2001), proximity
sensors (Clayton et al. 1989), laser scanners (Romero et al. 1997, Messerklinger and Springman
2007), digital image analysis (Macari et al. 1997, Sachan and Penumadu 2007), digital image
correlation (Bhandari et al. 2012), x-ray computed tomography (Desrues et al. 1996, Viggiani et
al. 2004), and photogrammetry (Kikkawa et al. 2006, Zhang et al. 2015a, Li et al. 2016). In
recent years, the popularity of photograph-based methods has surpassed non-photograph-based
methods due to their practicality, cost-effectiveness, and versatility. The limitations of the
photograph-based and non-photograph-based approaches were discussed in Salazar and Coffman
(2015a, 2015b), Salazar et al. (2015), and Salazar et al. (2017); the need for the use of
photogrammetry that relied upon internal cameras was presented.
Of the photograph-based triaxial monitoring examples in the literature (Macari et al.
1997, Alshibli and Sture 1999, Alshibli and Al-Hamdan 2001, Kikkawa et al. 2006, Gachet et al.
2007, Sachan and Penumadu 2007, Rechenmacher and Medina-Cetina 2007, Uchaipichat et al.
2011, Bhandari et al. 2012, Hormdee et al. 2014, Zhang et al. 2015a, Li et al. 2016, Salazar et al.
2017), only the techniques presented in Kikkawa et al. (2006), Zhang et al. (2015a) and Li et al.
(2016) utilized photogrammetry to obtain total and local volume changes of triaxial soil
specimens. The method presented in Zhang et al. (2015a, 2015b) and Li et al. (2016) involved
acquiring digital photographs of the specimen from outside of the cell wall. The photographs
were used to photogrammetrically reconstruct a digital, three-dimensional model. Due to the
refracted path of light between the specimen and the camera, computationally intensive
corrections were required to account for apparent distortion at the confining fluid-cell wall and
cell wall-atmosphere interfaces. This approach did appear to overcome previous limitations of
photograph-based measurement techniques, including Digital Image Analysis (DIA), Digital
![Page 116: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/116.jpg)
100
Image Correlation (DIC), and Particle Image Velocimetry (PIV). The Zhang et al. (2015a) and Li
et al. (2016) methods clearly present several advantages of a photogrammetric approach, but the
implementation of the method introduces additional processing complexity to account for optical
refraction and cell wall flexure (Zhang et al. 2015a, 2015b, Salazar and Coffman 2015b, Li et al.
2016).
As an alternative to the aforementioned methods, that utilized externally-acquired
photographs, a photogrammetric technique that utilized photographs that were captured from
within the triaxial cell was introduced (Salazar and Coffman 2015a, 2015b, 2016; Salazar et al.
2015, 2017). As described in Salazar and Coffman (2015a, 2015b, 2016) and Salazar et al.
(2015, 2017), small board cameras with pinhole apertures were mounted to diametrically
opposed towers that were located within the triaxial cell. The cameras were designed to
withstand exposure to the confining fluid (silicone oil) and the typical high confining pressures
associated with triaxial testing (up to 1,035 kPa). Despite the relatively wide field of view of
each camera (~70 degrees), the confined space within the triaxial cell (11.43 cm [4.5 in.] inside
diameter) required a total of ten camera devices (five devices stacked vertically on each tower) to
ensure full photographic coverage of a given soil specimen. The towers were mounted on a
guided track that allowed for rotation around the soil specimen as limited by the two top cap
drainage lines. Each time a set of photographs was captured, the drainage lines functioned as a
datum for camera stations by providing a consistent starting point for rotation. With the aid of
two pairs of magnets (located on the towers and outside of the cell), the towers were manually
rotated and stopped at prescribed intervals. Ten photographs were captured at each interval.
Photogrammetry software (PhotoModeler Scanner 2015) was then utilized to reconstruct the
surface for any soil specimen at any given stage during the triaxial testing.
![Page 117: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/117.jpg)
101
The primary advantage of the internal cell photogrammetry technique presented in
Salazar and Coffman (2015b, 2016) and Salazar et al. (2015, 2017) was direct observation of the
soil specimen during testing. The necessity to account for the refraction of light at the confining
fluid-cell wall and cell wall-atmosphere interfaces, as well as the curvature or deformation of the
cell wall, was therefore eliminated. Discussion of the method presented in Salazar et al. (2015)
was offered in Mehdizadeh et al. (2016), where it was claimed that yet another, simpler method
(Uchaipichat et al. 2011) could be employed to eliminate some of the cumbersome refraction
corrections. A closure to this discussion was provided in Salazar et al. (2017).
A comprehensive description of the steps used in the internal cell photogrammetry
approach is included herein. Furthermore, an evaluation of the accuracy of the approach, is
described. Three soil analog specimens were utilized for the evaluation. Specifically, an acrylic
specimen was used to verify the accuracy of the photogrammetric procedures. Additionally, a
brass specimen and a second acrylic specimen were used to examine the effect of the number of
photographs (ranging from 40 to 320 photographs) on the photogrammetric derivation of the
camera stations and on the determination of specimen volume. Undrained triaxial compression
and extension tests on kaolinite soil specimens are also described. The tests were performed to
demonstrate the viability of the internal photogrammetry approach to quantify total and local
deformations on the surface of the soil specimens during testing.
7.6. Evaluation of the Internal Photogrammetry Technique
As discussed herein, the performance of the internal cell photogrammetry approach that
was described in Salazar and Coffman (2015b, 2016) and Salazar et al. (2015, 2017) was
demonstrated by conducting a series of tests using soil analog specimens (brass and acrylic
specimens). Each step of the approach was evaluated prior to undrained triaxial compression and
![Page 118: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/118.jpg)
102
extension testing. These steps included the 1) calibration of each of the individual board cameras,
2) derivation of camera locations and orientations, 3) determination of optimal degree of rotation
between photograph-capturing intervals, 4) capture of photographs of the acrylic analog
specimen, 5) photogrammetric reconstruction of the acrylic analog specimen, 6) determination of
the volume of the acrylic analog specimen, and 7) evaluation of the accuracy of the volume
determination method. To illustrate the full evaluation process, a flow chart is presented (Figure
7.1). As a subset of Figure 7.1, the photogrammetric processes are further described in Figure
7.2.
Where DSLR is digital single lens reflex (camera), CTC is conventional triaxial compression, RTE
is reduced triaxial extension, V is volume, ΔV is change in volume, h is height, Δh is change in
height, d is diameter, Δd is change in diameter, εa is axial strain, εv is volumetric strain, and Af is
the area of the actual failure plane.
Figure 7.1. The process used to evaluate the internal cell photogrammetry technique and to
obtain test parameters.
See Table 7.2 See Tables 7.4, 7.5 See Figs. 7.8, 7.9
Volu
mes C
alc
ula
ted
3-D Scan
DSLR Camera Photogrammetry
Manual Measurements
Water Displacement
Comparison of Volumes
CTC and RTE Testing
of Soil Specimens
Capture photos of specimen with camera towers at 20° rotation interval
for any given stage of testing
DSLR Target Identification
Internal Camera Location/Orientation
Create 3-D volumes of specimen for any given stage of testing
Calculate specimen
V, ΔV, h, Δh, d, Δd, εa, εv, Af Visualize strains
Internal Cell Photogrammetry Evaluation
See Fig. 7.2 Volume Determination
Method Comparison
Internal Cell Photogrammetry
![Page 119: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/119.jpg)
103
Key
S1: Analog specimen (38.1 mm [1.5 in.] diameter by 76.2 mm [3.0 in.] length, nominal) with
targets used to derive location and orientation of internal cell cameras. Point cloud of targets was
fixed (as obtained from DSLR camera). Camera locations/orientations were fixed (as obtained
from the camera location/orientation step). S2: Larger analog specimen (44.5 mm [1.75 in.]
diameter by 88.9 mm [3.5 in.] length, nominal) with targets used to calculate locations of targets
on the specimen. Nomenclature: f is the focal length; w:h are the format size dimensions (width to
height ratio); x:y are the principal point coordinates; and K1, K2, K3, P1, P2 are dimensionless
distortion coefficients.
Figure 7.2. The process used to determine the volume of a specimen using internal cell
photogrammetry and to evaluate the sensitivity of photograph interval on the volume of the
specimen.
DSLR Target Identification
Internal Camera Location/Orientation
Target Identification
Capture photos of
coded calibration sheet
Create individual cell camera models
Outputs: f, w:h, x:y, K1, K2, K3, P1, P2
Create DSLR camera model
Outputs: f, w:h, x:y, K1, K2, K3, P1, P2
45° 30° 15° 5°
Capture photographs of S1 with camera towers at desired rotation interval
Capture photographs of S1 with DSLR camera
45° 30° 15° 5°
Derive camera locations and orientations
Create point cloud of coded targets located
on S1
Measure coded targets Output: physical
dimension with units
Identify/assign coded targets
Replace S1 photosets with corresponding
S2 photosets
Capture photos of S2 with camera
towers at desired rotation interval
Create point clouds of targets
located on S2
Identify/assign coded targets
Identify/assign coded target numbers on S2
Create 3-D volumes of S2
Capture photos of
coded calibration sheet
Compare volumes obtained from
45°,30°,15°, and 5° rotation intervals
See Table 7.3
Caliper
![Page 120: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/120.jpg)
104
7.6.1. Calibration of Board Cameras
Off-the-shelf software can be used to determine the intrinsic camera and lens parameters
that describe the interior orientation of a given camera. This camera calibration process is an
important step of the photogrammetric process when a high level of accuracy is desired. All
cameras involved with the processes described below, even the lensless board cameras as
originally described in Salazar and Coffman (2015a), were fully calibrated utilizing the single-
sheet calibration procedure, as outlined by Eos Systems, Inc. (PhotoModeler Scanner 2015).
Through this method, each of the ten cameras were positioned to capture convergent photographs
of a calibration grid from multiple camera stations and orientations (±90 degree roll angles),
with calibration targets well distributed throughout the photographs. These photographs were
then processed using the photogrammetric software, resulting in a calibration data file for each
camera. The data files contained intrinsic camera parameters that included the focal length (f),
the sensor format size (w:h), the principal point (x:y), and the dimensionless distortion
coefficients (K1, K2, K3, P1, P2), as reported in Table 7.1. The intrinsic camera parameters were
imported into all future photogrammetry projects that used the board camera-acquired
photographs.
![Page 121: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/121.jpg)
105
Table 7.1. Comparison of intrinsic camera parameters from calibration for all ten board
cameras utilized for internal photogrammetry.
Calibration Parameter
Camera Tower 1
Camera 1 Camera 2 Camera 3 Camera 4 Camera 5
Focal Length [mm] 3.3574 3.6599 3.3390 3.5936 3.5928
Format Size [mm] w: 5.145, h: 4.800 w: 5.149, h: 4.800 w: 5.150, h: 4.800 w: 5.152, h: 4.800 w: 5.148, h: 4.800
Princip. Point [mm] x: 2.573, y: 2.337 x: 2.659, y: 2.419 x: 2.427, y: 2.176 x: 2.806, y: 2.459 x: 2.782, y: 2.338
K1 3.87E-03 3.62E-03 2.37E-03 1.96E-03 2.32E-03
K2 -5.18E-05 -2.37E-04 3.89E-05 -7.22E-06 -4.96E-05
K3 0 0 0 0 0
P1 -1.96E-04 2.35E-04 1.36E-04 2.59E-04 1.72E-04
P2 2.50E-04 2.43E-04 -1.33E-04 4.49E-04 2.84E-04
Calibration Parameter
Camera Tower 2
Camera 1 Camera 2 Camera 3 Camera 4 Camera 5
Focal Length [mm] 3.4274 3.4498 3.2350 3.3464 3.5830
Format Size [mm] w: 5.147, h: 4.800 w: 5.150, h: 4.800 w: 5.153, h: 4.800 w: 5.154, h: 4.800 w: 5.150, h: 4.800
Princip. Point [mm] x: 2.823, y: 2.339 x: 2.608, y: 2.309 x: 2.423, y: 2.276 x: 2.397, y: 2.359 x: 2.370, y: 2.581
K1 4.27E-03 2.19E-03 2.75E-03 8.64E-04 2.03E-03
K2 -1.76E-04 -8.17E-05 -2.76E-05 -8.69E-05 2.05E-05
K3 0 0 0 0 0
P1 2.52E-04 1.14E-04 1.03E-04 9.82E-04 1.88E-05
P2 8.78E-05 3.78E-06 6.62E-05 -3.68E-04 -4.86E-05
Note: K1, K2, K3, P1, and P2 are dimensionless distortion coefficients.
7.6.2. Derivation of Camera Stations within the Triaxial Cell
With any photogrammetric project, it is necessary to obtain the exterior orientation
parameters (i.e. position and orientation) for each camera station used for capturing a
photograph. To derive this information for the board cameras that were internal to the triaxial
cell, the following procedure was performed. A cylindrical, brass analog specimen (38.1 mm [1.5
in.] diameter by 76.2 mm [3.0 in.] length, nominal) was wrapped with a sequence of black ringed
automatically detected (RAD) coded targets that were printed onto a sheet of white paper (to
provide contrast). The brass specimen was then placed upright on a flat surface. Additional RAD
targets were placed on the flat surface adjacent to the specimen to provide additional tie points
and improve the overall geometry and accuracy of the model. A fully calibrated, digital single
lens reflex (DSLR) camera with a fixed 28 mm lens were used to capture overlapping
![Page 122: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/122.jpg)
106
photographs from various positions around the brass specimen (approximately 40 photographs
total). A selection of these photographs were processed using the photogrammetric software,
which automatically identified and measured each RAD target. Several measurements acquired
using a caliper (between two distant targets) were input into the software program to define the
scale. Three-dimensional coordinates for all 286 RAD target locations were then exported as a
comma-delimited text file and used as control points in subsequent projects.
The same targeted brass specimen, which was used with the DSLR camera, was then
placed within the instrumented triaxial cell and a set of photographs was captured by the internal
board cameras at five-degree intervals. Although two 20-degree sections (40 of the total 360
degrees) were left with no directly perpendicular photographs, due to the presence of the two
diametrically opposed drain lines, the entire specimen surface was observed with the internal
board cameras. The set of five-degree interval photographs (total of 320) was processed using
the photogrammetry software. Visible RAD targets within the newly acquired set of photographs
were identified and assigned to the corresponding control point coordinates. With all control
points measured and photogrammetric processing complete, the locations (X, Y, Z) and
orientations (Omega, Phi, Kappa) for all 320 board camera stations were exported as a comma-
delimited text file. All future photograph acquisitions were assigned to the respective
photogrammetrically-derived camera locations and camera orientations listed in this file. These
steps served to establish a common 3D coordinate system for future photogrammetric
reconstructions.
7.6.3. Determination of Photograph-Capturing Intervals
Close-range photogrammetry of objects required not only full photographic coverage of
all specimen surfaces, but overlapping photographs such that all points to be measured were
![Page 123: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/123.jpg)
107
clearly visible in at least two but ideally three or more photographs. To meet this requirement,
photographs were captured at specified intervals as the towers were manually rotated around the
specimen. It was desired to minimize the number of photographs required to reconstruct the
specimen, while maintaining a high degree of accuracy and precision. By increasing the base line
distance (and therefore rotation angle) between camera stations, the accuracy of measurements in
object space (i.e. on the specimen surface) was expected to increase. In turn, increasing the base
line distance reduced 1) the number of photographs and 2) the available overlap between
adjacent photographs. It was therefore desired to determine the optimal intervals at which
photographs should be captured. A sensitivity study was performed to determine the ideal angle
between adjacent sets of photographs. The study was conducted by placing a different analog
specimen (acrylic, 44.5 mm [1.75 in.] diameter by 88.9 mm [3.5 in.] length, nominal) into the
instrumented triaxial cell and capturing photographs of the specimen at five degree intervals (320
photographs, total). The larger specimen was selected because it represented the maximum
dimensions that would be achieved during large-strain triaxial compression tests (maximum
diameter) or extension tests (maximum height) on actual soil specimens. The cell remained
empty (air, instead of confining fluid) for this stage of the evaluation process. The sensitivity of
the camera stations to the angle between the photograph capturing intervals was evaluated for
45-, 30-, 15-, and five-degree intervals, which corresponded to 40, 60, 110, and 320 photographs,
respectively. These intervals were chosen because each interval was divisible by the next,
allowing for one common photoset to be used.
7.6.4. Capture of Photographs to Determine Accuracy in Confining Fluid
The same procedures that were utilized to 1) derive the board camera stations using the
brass analog specimen (in air) and to 2) determine the ideal angle between photos using the
![Page 124: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/124.jpg)
108
large, acrylic analog specimen (also in air) were employed to evaluate the method in confining
fluid. A second, smaller acrylic analog specimen (38.1 mm [1.5 in.] diameter by 76.2 mm [3.0
in.] length, nominal) was submerged in confining fluid (silicone oil) within the triaxial cell for
the procedure. RAD-coded targets were printed on a sheet of temporary tattoo adhesive paper
and adhered to the surface of the acrylic specimen. The photogrammetry procedures that were
previously outlined in Section 7.6.2 were repeated and final camera stations were derived for all
future tests performed with the confining fluid-filled cell. This was necessary to ensure accuracy
of the camera stations when the cameras were submerged in the confining fluid. The coded
targets were removed and a different sequence of coded targets was adhered to the surface of
subsequent specimens. The new targets were used because it distinguished them from the targets
that were already identified to create the control point cloud (used to derive the camera stations).
The acrylic specimen was then placed within the triaxial cell filled with confining fluid once
more and photographs were captured to reconstruct the specimen. This second set of photographs
of the acrylic specimen was necessary because it would not have been a fair assessment to derive
the target locations on the surface of the specimen using the same photographs that were utilized
to derive the camera stations.
7.6.5. Photogrammetric Reconstruction of a Specimen
The photographs of the two acrylic analog specimens (large specimen used to evaluate
the sensitivity of the photograph-capturing interval and smaller specimen used to evaluate the
accuracy of the technique when subjected to the confining fluid) that were captured from within
the triaxial cell were processed within software to photogrammetrically reconstruct the
specimens. The photogrammetry projects that were created during the camera location and
orientation step, to establish the 3D coordinate system, were modified by replacing the
![Page 125: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/125.jpg)
109
photographs within the existing projects with the newly acquired photographs of the acrylic
specimens. This ensured that the already derived camera stations (location and orientation)
remained constant, thereby enabling the greatest possible accuracy for the close-range
photogrammetry technique. Each visible target on the surface of the acrylic specimens was then
identified and measured in at least three photographs and assigned to the respective unique
identification number (384 and 283 total targets for the large- and small-acrylic specimens,
respectively). Once a target was measured in three or more photographs, three-dimensional
coordinates for that point were computed and reported by the software. The circular centers of
the targets provided a reliable means of identifying the precise locations of the targets. To aid in
the reliable identification of common points on the ends of the specimen, high contrast markers
were added to the porous stones on both ends of the specimen. The intersections between the
markers, the porous stones, and the ends of the specimen served to identify common points along
the ends of the specimen. Internal quality feedback within the software aided in identifying and
reducing point measurement errors, thereby 1) ensuring the quality of the photogrammetry
projects and 2) providing consistency among each of the projects that were processed. The
quality feedback metrics included total error, residuals, and point precision values.
After all of the points on the surfaces of the specimens were identified, radial curves were
drawn through the 3D points on the surface of the virtual specimens. Surface tools were utilized
to create outward-facing surfaces on the specimens; these surfaces were created by using the
curves as the edges of each surface, and to cap the open ends of the specimens. The virtual
specimens therefore took shape using the newly created surfaces; however, the photogrammetry
software did not correctly calculate the internal volumes of the virtual specimens, nor were the
surfaces “watertight”. The 3D models were therefore exported in a wavefront format (.obj
![Page 126: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/126.jpg)
110
extension) to allow for further analysis using a software program that was more suited to
determining the accurate volume of a virtual object. A 3D computer-aided drawing (CAD)
software (Geomagic Design X 2015) was utilized for this purpose.
7.6.6. Determination of a Specimen Volume
Each 3D model exported from the photogrammetry software consisted of a number of
disconnected polygonal bands wrapped transversely around the surface of the model. Narrow
gaps between these polygonal bands were sealed using the global remesh and healing tools
within the software. The remesh tool worked by essentially shrink-wrapping the 3D model with a
new, improved surface that was free of holes, slivers, and other topologic imperfections. The
used software tools are proprietary but the results were similar to what would be expected from a
Poisson surface reconstruction, like that described by Kazhdan and Hoppe (2014). The settings
for this tool were adjusted so that the number of polygons that made up the output model was
100 times the number of polygons of the input model. The increase in the quantity of polygons
reduced the potential for rounding that was observed along sharp edges. Moreover, the healing
tool was then used to detect and remove any small clusters of free-floating polygons that were
not actually part of the surface of the models. After the final watertight models were created, the
calculation of the volume of each model was revealed when selecting on the properties of the
model. The exact method used to calculate volume by the software was not reported by the
software publishers, but the used method was likely similar to the process described by Mirtich
(1996).
7.6.7. Accuracy of Technique
To evaluate the accuracy of the internal cell photogrammetry approach, that is presented
herein, several other techniques were also employed to determine the volume of the smaller
![Page 127: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/127.jpg)
111
acrylic analog specimen. The techniques included 1) the aforementioned internal
photogrammetry technique, 2) photogrammetry using DSLR camera obtained photographs only,
3) a 3D scanning technique, 4) manual measurements using a caliper and pi tape, and 5) a water-
displacement technique. Each of these not previously mentioned techniques (techniques 2-5) are
described in this section. Based on a review of the literature, no universal method exists to
evaluate the absolute or “true” accuracy of a volume determination technique. The amount of
difference relative to an external reference, often termed “error”, is only meaningful when the
nature of the external reference is reported. To provide a metric for comparison between the
volumes of the smaller acrylic specimen, as obtained using each technique, the difference was
evaluated relative to the water displacement technique. This technique was selected, because it
was based on well-established procedures documented in ASTM Standard D698 (ASTM 2012)
to determine the interior volume of a Proctor mold.
7.6.7.1. DSLR Camera Photogrammetry
For the DSLR camera survey technique, the smaller acrylic specimen was placed on a
table and approximately 40 photographs were captured of the specimen from various angles. The
photographs were imported into the photogrammetry software and a selection of the best photos
were processed. Common points (coded targets), on the surface of the specimen, were identified
and referenced to ensure that the points appeared in at least three photos. Measurements were
also imported to define the scale (known distance between select points). In a similar fashion to
the internal photogrammetry technique, surfaces were created on the virtual specimen and the
model was exported for processing and analyzed within the 3D CAD software.
![Page 128: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/128.jpg)
112
7.6.7.2. 3D Scanning
By definition, 3D scanning is the use of a specialized instrument to rapidly record the 3D
information of an object or environment. A close-range 3D digitizing system (Breuckmann
SmartScan3D HE) that utilized fringe projection, or structured white light technology, was
employed to obtain the 3D data of the acrylic specimen. Specifically, a projector, two 5-
Megapixel color cameras, and multiple lenses were utilized to facilitate the 3D measurements. A
series of patterns (or fringes) were cast onto the specimen and the difference in the pattern from
each camera was utilized to compute a series of discrete measurements or 3D points. The
instrument captured approximately 150,000 points per individual scan.
The smaller acrylic specimen was scanned with a set of M-125 lenses (i.e. 125 mm
diagonal field of view at the optimal working distance of one meter). The M-125 lenses, the
highest resolution lenses available for this scanner, were used to achieve the highest possible
spatial resolution of approximately 60 μm horizontal. To begin the process of scanning, the
instrument was calibrated using 1) the prescribed procedure that was recommended by the
manufacturer, 2) a set of calibration targets, and 3) the companion 3D digitizing software
(OPTOCAT 2013R2). The calibration procedure reported an average accuracy of object points
of 15.41 μm in the X, 0.74 m in the Y, and 26.75 μm in the Z dimension (depth from scanner).
The specimen was made of an acrylic material that was partially transparent. To prevent scan
errors caused by light scattering during fringe projection, a thin coat of matte white spray paint
was applied to the specimen. Several spherical adhesive targets were also placed on each side of
the specimen to aid in the scan-to-scan alignment procedures during data processing. The
specimen was then placed at a 45-degree angle on an automated turntable (Figure 7.3) and
scanned at 20-degree intervals for a total of 18 scans. Two other manually positioned scans were
![Page 129: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/129.jpg)
113
collected to fill in areas not visible during the turntable rotations. All of these data (20 scans)
were then processed using the digitizing software. The basic processing steps that were
performed included: 1) an iterative global best-fit alignment of all scans, 2) overlap reduction to
remove scan data collected at a high angle of incidence, 3) merging of individual scans to create
a single polygonal mesh, 4) smoothing to remove small amounts of noise and other scan
artifacts, and 5) hole-filling using the semi-automated tools that were available. The final 3D
model, as presented in Figure 7.3, was composed of approximately 685,000 polygonal faces and
approximately 343,000 vertices.
Figure 7.3. (a) Photograph of, and (b) three-dimensional, watertight model of small-acrylic
analog specimen with spherical adhesive targets (removed during processing), as obtained
during 3D scanning of specimen.
7.6.7.3. Manual Measurements
For the method that consisted of manual measurements, a linear caliper (with a resolution
of 0.05 mm) was used to measure the length of the acrylic specimen (average of three
measurements) and a pi tape (with a resolution of 0.01 mm) was used to measure the diameter of
the specimen (average of three measurements). The volume of the specimen was then calculated
based on the average measurements.
(a) (b)
![Page 130: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/130.jpg)
114
7.6.7.4. Water Displacement
The same procedures that are commonly utilized to measure the volume of a Proctor
mold, as outlined in ASTM Standard D698 (ASTM 2012), were used to measure the volume of
the specimen. Specifically, after the volume of a Proctor mold was determined using the water-
filling method that is described in the Annex of the ASTM Standard, the specimen was placed
into the Proctor mold and submerged in de-ionized and de-aired water to determine the amount
of water that was displaced by the specimen. The mass of the acrylic specimen was determined
before and after water submersion to ensure that no water was imbibed by the specimen during
the testing.
7.6.8. Limitations and Sources of Error
The limitations of, and the identified sources of error associated with, the described
photogrammetry technique are discussed herein. The identified systematic errors were mitigated
during the process of collecting, processing, and evaluating data. Additionally, a schematic of the
relevant qualitative factors that influence the accuracy of photogrammetry applications is
presented as Figure 7.4.
![Page 131: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/131.jpg)
115
Shading highlights the characteristics of the photogrammetry methodology presented in this study.
Figure 7.4. Qualitative factors affecting accuracy in photogrammetry (modified from Eos
Systems, Inc. 2015).
7.6.8.1. Precision of Repeat Interval Stops
Measured from a fixed starting position (in contact with the drainage lines), the camera
towers stopped at prescribed rotation intervals around the specimen to allow for repeat
occupation (during a given photogrammetry project and between successive photogrammetry
projects). The method relied upon the capture of photographs from the exact same locations with
each repetition, because photographs with known (derived) camera stations were replaced with
new photographs (thereby assigning the derived locations and orientations to the new
photographs). Although the same locations were reoccupied for each test, any deviation from the
photogrammetrically derived location could have resulted in error in the three-dimensional
coordinate of an observed point within the replaced photographs.
No
calibrationLess than
15 degrees
Few targets
per photo,
low
coverage
No targets,
all user
marked
Points
mostly on
only two
photos
Photo
RedundancyTargets
Camera
Resolution
Camera
Calibration
Method
Angles
between
Photos
Photo
Orientation
Quality
Expected
Accuracy
Lowest
Accuracy
Highest
Accuracy
11
Megapixel
5-6
MegapixelAverage
Accuracy
Video
640x480
Field
calibrated
camera
Most targets
on 8 or more
photos
35+ targets
per photo,
50-80%
coverage
Between 60
and 90
degrees
Some
naturally lit
targets
Retro-
reflective
Many good
quality
naturally lit
Calibrated
camera
Inverse
camera
All points on
3+ photos
15+ targets
per photo,
25- 60%
coverage
Between 15
and 60
degrees
![Page 132: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/132.jpg)
116
7.6.8.2. Model Refinement
The number of targets that were utilized limited the mesh refinement of the surface of
each specimen. Furthermore, the number of targets that were utilized was related to processing
time and to the minimum size of targets. To maintain the automated target identification
capability of the photogrammetry software, a target center diameter of at least 30 pixels was
utilized. This resulted in the use of 286 targets that were evenly distributed (center to center
spacing of 5.65 mm) across the surfaces of the 38.1 mm (1.5 in.) in diameter by 76.2 mm (3.0
in.) in length (nominal) brass and acrylic soil specimens.
7.6.8.3. External Geometry Measurements
To scale a photogrammetry project, one or more external reference measurements was
required to be input. These reference measurements were in the form of a known distance
between two measured points located within the project. The resulting overall accuracy of a
project was therefore limited to the accuracy of the input measurements. To mitigate the impact
of this source of error, multiple reference measurements were made for various target pairs
within the project.
7.6.8.4. Determination of Specimen Ends
The most difficult aspect of processing the photographs of a specimen was the reliable
identification of the ends of the specimen (i.e. picking points along the edges at the two ends of
the specimen). Picking end points was challenging because distinct markers had to be identified
subjectively in adjacent photographs without the help of target centers. This challenge has often
been understated or not discussed in the literature, but should not be overlooked. To aid in the
reliable identification of specimen ends, high contrast markers were applied to the porous stones
on the ends of the specimens.
![Page 133: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/133.jpg)
117
7.7. Implementation of the Internal Photogrammetry Technique for Soil Specimens
Two undrained triaxial tests were performed on kaolinite soil specimens to assess the
viability of determining: 1) total and local strains, 2) total volume and volume changes at any
given stage of testing, and 3) the actual failure plane of a soil specimen. Specifically, one
undrained, conventional triaxial compression (CTC) test and one undrained, reduced triaxial
extension (RTE) test were conducted. The tests were performed with advanced, automated
triaxial apparatus that included pore water pressure and pore water volume measurements. In a
typical triaxial test, the exact total specimen volume at any given stage of testing (prior to
consolidation, prior to shearing, or during shearing), must be back-calculated from testing and
post-testing data using phase relationships and assumptions (most notably the right circular
cylinder assumption, see ASTM D4767-11, 2011). To illustrate this concept, a schematic of the
stages of a typical triaxial compression test is presented as Figure 7.5. This method of calculating
specimen volume often leads to erroneous results without any means of verification. The
implemented photogrammetry technique provided a means to independently determine the
volume of a soil specimen, during any desired stage of testing, without the need to rely upon
erroneous assumptions made during back-calculation.
![Page 134: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/134.jpg)
118
1. Pre-test: Mass (m) and water content (w), measured; Volume (V) calculated using caliper measurements. 2. Back-pressure saturation: Drain lines filled. Total volume change (ΔV) from pore pump measurements. This volume change includes air 1) purged from lines, and 2) going into suspension. 3. K0 Consolidation: Sample ΔV from pore pump measurements. 4. Shearing: m, w, and V assumed to be equal to post-test m, w, and V (if undrained); calculated from pore pump measurements (if drained). 5. Post-test: m and w, measured. Shear strength determined based on corrected area (Ac).
Figure 7.5. Typical measurements and calculations required to determine the phase
diagram of the soil specimen during a conventional triaxial compression test.
The soil specimens consisted of commercially available kaolinite soil (Kaowhite-S)
obtained from the Thiele Company (Sandersonville, Georgia). The specimens were slurry-
consolidated in an acrylic consolidometer under an overburden stress of 138 kPa (20 psi).
Specimens with nominal dimensions of 7.62 cm in length and 3.81 cm in diameter were extracted
from the consolidation apparatus and weighed. Using temporary tattoo paper, RAD-coded targets
were applied to the surface of the first membrane surrounding the sample. The membrane was then
placed onto the specimen, and a second membrane was applied over the first membrane (to reduce
the potential for liquid transfer or gas permeation). During the specimen preparation phase, care
was taken to minimize the amount of disturbance on the soil specimen. The confining cell was
filled with silicone oil, the top and bottom drain lines to the specimen were flushed to remove air
from the lines, and the specimen was back pressure saturated (B-check equal to 0.95 or higher)
before proceeding to the consolidation phase. During each test, the specimen was consolidated
under K0-conditions to a vertical effective stress of 310 kPa (45 psi). During consolidation, the
![Page 135: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/135.jpg)
119
changes in pore pump volume measurements were used to determine changes in pore water volume
of the specimen. Upon completion of consolidation, the drain lines were closed and the specimen
was sheared under undrained conditions (strain rate of 0.5 percent per hour). For the CTC test, the
shearing was paused at intervals of 0, 2, 4, 6, 8, 11.5, and 15 percent axial strain. At each of these
strain intervals, the ten board cameras were used to capture photographs of the specimen at 20-
degree rotation intervals (total of 80 photographs per strain interval), as further discussed in
Section 7.8.2. Similarly, for the RTE test, the shearing was paused at intervals of -0, -2, -4, -6, -8,
-10, -12, -15 percent axial strain and photographs of the specimen were captured. For each strain
interval, the test was paused for less than 30 minutes. For completeness, a photograph of the
instrumented triaxial cell, as utilized in the RTE test, is presented (Fig. 7.6).
Processing procedures were identical to those employed to model the acrylic analog
specimen. After 3D models of the soil specimens were exported to wavefront format files, the
models were further analyzed within 3D CAD software. Local displacements on the surface of
each soil specimen were visualized using the built-in mesh deviation function. The software tool
likely functioned similar to the process described by Cignoni et al. (1998). Utilization of this
function allowed for watertight meshes to be overlayed (with a common coordinate system) to
compare the positive or negative changes between the surfaces of the two meshes. The post-
consolidation mesh was selected as the reference mesh for comparisons. A color-graded scale
was selected to visualize the magnitude of changes (cooler colors corresponded to negative
changes while warmer colors correlated to positive changes).
![Page 136: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/136.jpg)
120
Figure 7.6. Photograph of the kaolinite specimen within the photogrammetrically
instrumented triaxial cell during the shearing stage of the extension test.
In addition to the undrained triaxial compression and triaxial extension tests, one
additional unconfined compression (UC) test was performed. The purpose of the UC test was to
compare 1) the calculated volumes during a test within the triaxial cell by utilizing the internal
photogrammetry technique, with 2) the calculated volumes during a test outside of the triaxial
cell utilizing the DSLR camera photogrammetry technique. The soil specimen was prepared in
an identical way to those specimens that were used in the two triaxial tests. RAD-coded targets
1. Load frame reaction rod (two quantity)
2. Cable with nine-pin feed-through connector
(four pins for the internal load cell, two pins
for the switchboard power supply, one pin for
the video signal, two pins unused)
3. Uplift prevention rod (two quantity, for
extension testing only)
4. Piston housing
5. Piston lock
6. Vacuum line for acrylic top cap vacuum
connection
7. Top platen of cell
8. Fastening rod (three quantity)
9. Piston
10. Switchboard for camera timing (as shown in
Salazar et al. [2015])
11. Electrical jumpers for individual camera power
supply (red)
12. Internal load cell
13. Electrical jumpers for common grounding and
video signals (green and yellow, respectively)
14. Acrylic top cap (triaxial extension vacuum cap
shown)
15. Drain line (connection to top cap)
16. Pore pressure transducer
17. Camera tower (two quantity, 5 cameras each)
18. Soil specimen within membrane (RAD-coded
targets adhered to membrane)
19. Rotating Delrin® bearing track
20. Top drain line and drain valve (black)
21. Bottom drain line and drain valve (red)
22. Cell pressure application line
1
2
3 4 5
6
7
12
15
16
17
18
22
14
19 20
21
13
11
10
9 8
![Page 137: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/137.jpg)
121
were applied to the surface of the membrane and additional targets were placed on the loading
frame around the specimen to provide tie points for photogrammetric processing. The specimen
was sheared under unconfined conditions (although the specimen was wrapped in a membrane)
at a strain rate of 0.5 percent per hour. During the test, the shearing was paused at intervals of 0,
2, 4, 6, 8, 11.5, and 15 percent axial strain and approximately 40 photographs of the specimen
were captured at each strain interval. During the processing phase, the best 12 photos of the 40
photos that were captured for each strain interval, were selected and processed so that targets on
the surface of the specimen appeared in at least three photographs. Following the same
procedures as those used for the internal photogrammetry technique, 3D models were created
within photogrammetry software and were exported for further analysis within the 3D CAD
software.
7.8. Results and Discussion
The results from the evaluation of the internal cell photogrammetry technique are
presented herein. Furthermore, a discussion of the amount of error associated with the technique
and the sensitivity of the photograph-capturing interval are presented. The accuracy of the
utilized photogrammetry technique is discussed and the limitations are highlighted.
7.8.1. Volume Comparisons
The differences between the volumes determined using the various measurement
techniques (internal photogrammetry, DSLR photogrammetry, 3D scanning, manual
measurements) relative to the arbitrary reference technique (water displacement) fell within one-
half of one percent, as presented in Table 7.2. There was good agreement between the volumes
determined from the internal photogrammetry technique and reference technique (0.13 percent
difference). These difference values were expected to be greater for the techniques presented
![Page 138: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/138.jpg)
122
herein than the difference values reported in the literature. This was expected because of the
relatively small size of the specimen that was utilized for evaluation of the internal
photogrammetry technique (nominal dimensions of 7.62 cm in length and 3.81 cm in diameter),
as compared to larger size specimens contained within the literature (typically, 10.16 cm in
length and 5.08 cm in diameter, or 14.22 cm in length and 7.11 cm in diameter). Despite the
increased sensitivity to volume determination error for the specimen used in this study, the
volume differences for all five measurement techniques were small (≤ 0.50 percent). The smaller
specimen size was utilized because of the reduced drainage distance, which significantly reduced
the time required for the completion of the consolidation phase of testing.
Based on a variety of tests conducted (prior to results reported in this study), the
repeatability of determining the volume of an analog specimen fell within 0.011 percent for the
3D scanning technique and within 0.084 percent for the internal photogrammetry technique.
Although the repeatability of the DSLR camera photogrammetry technique was not studied, it is
expected to be comparable to the repeatability of the internal photogrammetry technique.
Table 7.2. Comparison of small-acrylic analog specimen volumes as obtained using five
different techniques.
Volume Determination Method
Volume of Specimen [cm3] Mean [cm3]
Difference from Reference [%] Repetition
1 2 3
Water Displacement 94.97 95.60 95.47 95.35 Reference
Manual Measurements 95.82 95.82 95.82 95.82 0.50
3-D Scan 95.64 - - 95.64 0.31
DSLR Photogrammetry 95.62 - - 95.62 0.29
Internal Photogrammetry 95.22 - - 95.22 -0.13
![Page 139: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/139.jpg)
123
7.8.2. Photograph Interval
Although it appeared that the repeatability of derived camera locations was sensitive to
the photograph interval (degree of separation between sets of photographs), as indicated by
convergence of camera locations in Figure 7.7, the effect was considered negligible (within
0.045 pixels for the maximum difference in camera location). The relationship between derived
camera orientation and photograph interval was not directly meaningful. Therefore, the influence
of the tower rotation interval on the determination of specimen volume was examined (Table
7.3). For the volume (as calculated from four photogrammetric reconstructions, using 45-, 30-,
15-, and five-degree photograph intervals), the standard deviation was equal to 0.34 cm3, and the
range was equal to 0.70 cm3. The determination of volume was therefore not sensitive to the
photograph interval. Thus, to 1) match the 20-degree gaps surrounding the drainage lines within
the triaxial cell, thereby providing consistent photograph intervals, and 2) maintain sufficient
overlap between photographs, an interval of 20 degrees was selected. This resulted in 80
photographs and approximately 240 minutes of processing time per test.
Table 7.3. Comparison of large-acrylic analog specimen volumes as determined during
internal photograph interval sensitivity test.
Rotation Interval
[Degrees]
Number of Photos
Computational Cost [minutes]
Specimen Volume
VT , [cm3] Summary Statistics
45 40 120 135.17 Mean Volume [cm3] 135.56
30 60 180 135.37 Standard Deviation [cm3] 0.34
15 110 330 135.87 Standard Error [cm3] 0.17
5 320 960 135.80 Coefficient of Variation [%] 0.25 Range [cm3] 0.70
Note: Photographs acquired using internal board cameras.
![Page 140: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/140.jpg)
124
Figure 7.7. Sensitivity of derived camera location difference to interval between
photographs.
7.8.3. Testing of Internal Photogrammetry System on Soil Specimens
The volume of the various soil specimens was determined at numerous levels of axial
strain during both the CTC and RTE tests, as well as during the UC test. The CTC and RTE tests
were performed in an undrained condition and therefore the total volume of the specimen was
not expected to change during the shearing phase of each test. Likewise, the UC test was
undrained. The volumes that were measured during each test, and the summary statistics for each
test, supported this hypothesis. The results from the CTC test are presented in Table 7.4. The
volume change during the consolidation phase was determined to be 6.56 cm3, using the internal
photogrammetry technique. As a comparison, the volume change determined from the pore
pump measurements was equal to 6.81 cm3 (temperature corrected) and the change calculated
from the displacement transducer measurements was equal to 6.70 cm3 (using the assumption
that the cross-sectional area of the specimen remained constant during K0 consolidation). The
0
0.01
0.02
0.03
0.04
0.05
0153045
Ca
me
ra L
oc
ati
on
Dif
fere
nc
e [
Pix
els
]
Photograph Interval [Degrees]
Average Difference
Maximum Difference
20
0.012
![Page 141: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/141.jpg)
125
internal photogrammetry approach therefore underpredicted the volume change by 3.7 percent,
as compared to the pore pump measurements, and by 2.1 percent, as compared to calculations
using the change in specimen height.
The results from the RTE test and the UC test are presented in Table 7.5 and Table 7.6,
respectively. For the CTC, RTE, and UC tests, the small changes in total volume, during
undrained shearing, were likely a result of the sensitivity to limited refinement of the 3D model
surface (function of the number of targets on the membrane). As indicated by the standard
deviation of total volumes calculated during the CTC test (0.37 cm3), as compared to the
standard deviation during the RTE test (0.27 cm3), the variability was greater for the CTC test.
The likely cause of the greater variability during the CTC test was that the target refinement was
more sensitive to the local deformations on the surface of the specimen during compression
(uneven bulging) than during extension (fairly uniform necking). Comparison with the results
from the UC test (standard deviation of 0.69 cm3) revealed that even with the high resolution
DSLR camera photogrammetry technique there was variability in the volumes, further
supporting the hypothesis that the model refinement (number and density of targets on surface of
the specimen) affected the accurate determination of specimen volume throughout a test.
![Page 142: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/142.jpg)
126
Table 7.4. Volumes of kaolinite soil specimen as determined throughout the triaxial
compression test and corresponding summary statistics.
Testing Phase
Axial Strain εa , [%]
Volume VT , [cm3]
Summary Statistics
Consolidation Pre-consolidation 89.72 Change in Volume During
Consolidation [cm3] 6.56
0 83.16
Shear
2 82.92 Mean Volume During Shear [cm3]
83.37 4 83.28
6 83.27 Standard Deviation [cm3] 0.37
8 83.28 Standard Error [cm3] 0.14
11.5 84.10 Coefficient of Variation [%] 0.45
15 83.55 Range [cm3] 1.18
Note: Photographs acquired using internal board cameras.
Table 7.5. Volumes of kaolinite soil specimen as determined throughout the triaxial
extension test and corresponding summary statistics.
Testing Phase
Axial Strain εa , [%]
Volume VT , [cm3]
Summary Statistics
Shear
0 79.88 Mean Volume 80.30
8 80.40 During Shear [cm3]
10 80.32 Standard Deviation [cm3] 0.27
12 80.28 Standard Error [cm3] 0.12
15 80.64 Coefficient of Variation [%] 0.34
Range [cm3] 0.76
Note: Photographs acquired using internal board cameras.
Table 7.6. Volumes of kaolinite soil specimen as determined throughout the unconfined
compression test and corresponding summary statistics.
Testing Phase
Axial Strain εa , [%]
Volume VT , [cm3]
Summary Statistics
Shear
0 91.01 Mean Volume 91.35
2 91.46 During Shear [cm3]
4 90.99 Standard Deviation [cm3] 0.69
6 90.90 Standard Error [cm3] 0.26
8 90.75 Coefficient of Variation [%] 0.75
11.5 91.57 Range [cm3] 2.00
15 92.75
Note: Photographs acquired using DSLR camera.
![Page 143: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/143.jpg)
127
The localized displacements of each specimen were visualized qualitatively for the CTC
and RTE tests. Specifically, the displacements were visualized for the consolidation phase of
testing, as presented in Figure 7.8, and for the shearing phase, as presented in Figure 7.9. During
the consolidation phase, the small strains in the radial direction of the specimen were somewhat
unexpected, as the triaxial testing apparatus was programmed for K0-consolidation by which the
diameter of the specimen should have remained constant throughout the consolidation phase of
the test. In the CTC test (Figure 7.9a), the actual failure plane of the soil specimen was evident
from the shear banding behavior at larger strains (greater than eight percent axial strain).
Conversely, necking behavior was observed for the specimen in the RTE test (Figure 7.9b).
Additional test data, including stress-strain and excess pore pressure relationships, were reported
in Salazar et al. (2017).
Figure 7.8. Visualization of photogrammetry-obtained, three-dimensional models of
kaolinite specimen during K0-consolidation phase of triaxial test (warm colors indicate
positive deformation and cool colors indicate negative deformation).
Pre-consolidation Post-consolidation
+2.175 mm -2.175 mm
![Page 144: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/144.jpg)
128
Note: Photographs on the right are of post-test, oven-dried specimens.
Figure 7.9. Visualization of photogrammetry-obtained, three-dimensional models of
kaolinite test specimen during (a) conventional triaxial compression, and (b) reduced
triaxial extension tests up to 15 percent axial strain during shearing (warm colors indicate
positive deformation and cool colors indicate negative deformation).
Although the application of the internal photogrammetry technique required
modifications to a standard triaxial testing cell, the components of the system that were presented
were inexpensive, especially when compared with the cost of the triaxial testing equipment.
Furthermore, the technology associated with the modifications was not complex (Salazar and
Coffman 2015a, 2016 and Salazar et al. 2015). With the reduced computational cost of using less
photographs to reconstruct a given test specimen, the added data collection and processing time
was insignificant (hours) when compared with the total time (sample preparation, setup, testing,
Immediately Post-consolidated
εa = 0% εa = 2% εa = 8% εa = 15% εa = 6%
εa = -8% εa = -12% εa = -15% εa = -10%
(a)
(b)
Immediately Post-consolidated
εa = 0%
+2.175 mm -2.175 mm
![Page 145: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/145.jpg)
129
and data reduction) for a typical triaxial test (weeks). Furthermore, the technique was no more
computationally costly than other photogrammetry methods found in the literature.
7.9. Conclusions
The internal cell photogrammetry technique was utilized to determine the volume of soil
specimens during all stages of undrained triaxial compression (CTC) and triaxial extension
(RTE) tests. The camera instrumentation, internal to the triaxial cell wall, allowed for direct
observation of the entire surface of the soil specimens throughout the triaxial tests and avoided
the need to correct for refraction and cell wall flexure involved with other methods in the
literature. The principles of close-range photogrammetry were utilized to enable accurate 3D
reconstructions of the soil specimens. Prior to triaxial testing, a variety of outside-of-cell volume
determination techniques, including DSLR camera photogrammetry, 3D scanning, manual
measurements, and water displacement techniques were employed to provide comparisons with
the volume of an acrylic analog specimen as determined utilizing the internal cell
photogrammetry technique. Results from the internal photogrammetry technique fell within 0.13
percent of the reference technique (water displacement) and results from all comparison
techniques fell within 0.50 percent. To minimize processing time to approximately 240 minutes,
a balance was struck between the number of photographs utilized (80 photographs total, 10
photographs captured at each 20 degree interval around a specimen) and the reliability in
photogrammetric measurements. 3D models were produced using commercially available
software and localized displacements that developed during the triaxial testing were visualized
and reported.
![Page 146: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/146.jpg)
130
7.9.1. Potential Applications and Future Improvements
There are several potential applications for using the internal photogrammetry technique.
The approach may be utilized to provide verification of axial and radial strain measurements at
any point on the surface of the specimen or at the end cap connections. Furthermore, the strain-
based approach could be used in conjunction with 3D finite element analysis techniques to
predict the stress distribution throughout the specimen. This inverse solution will aid in
developing understanding into the constitutive models used to predict soil behavior.
Although other techniques may be used to measure total pore volume changes during
consolidation (i.e. pumps), these techniques may not necessarily be used to measure total volume
changes during shearing, nor will localized volume changes be captured. The technique
presented in this work may therefore be employed to monitor both total and local volume
changes during drained and undrained triaxial tests.
Future improvements to the internal photogrammetry system may facilitate increased
accuracy of the results. A higher degree of precision in the reoccupation of photograph interval
stops around the specimen would reduce error arising from the predetermined camera stations.
Therefore, a mechanized rotating track base is recommended for future applications.
Furthermore, future projects may also incorporate a geometric constraint that allows some small
amount of deviation from the known camera positions, but only along a modeled arc
representing the circular path of the camera track.
To increase the level of refinement on the surface of a specimen, a greater number of
targets may be required. However, the size of (and therefore the number of) the targets that were
utilized was limited, due to the resolution of the modified board camera devices. To reduce the
approximations between targets, improved camera resolution will allow for denser target
![Page 147: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/147.jpg)
131
coverage on the specimen surface. Furthermore, improvements in automatic target identification
algorithms will result in reduced time required for processing.
7.10. Acknowledgements
This material is based upon work supported by the National Science Foundation Graduate
Research Fellowship under Grant No. DGE-1450079, as awarded to S.E. Salazar. The authors
would also like to acknowledge 1) support from the University of Arkansas Provost
Collaborative Research Grant awarded to R.A. Coffman (PI), M.L. Bernhardt (Co-PI) and A.
Barnes (Co-PI) and 2) undergraduate student support from the United States Department of
Transportation (USDOT) Office of the Assistant Secretary for Research and Technology (OST-
R) under Research and Innovation Technology Administration (RITA) Cooperative Agreement
Award No. OASRTRS-14-H-UARK awarded to R.A. Coffman (PI) and T. Oommen (Co-PI).
The views, opinions, findings, and conclusions reflected in this manuscript are the responsibility
of the authors only and do not represent the official policy or position of the USDOT/OST-R, or
any state or other entity.
![Page 148: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/148.jpg)
132
7.11. References
Alshibli, K. A. and Sture, S., 1999, “Sand Shear Band Thickness Measurements by Digital
Imaging Techniques,” J. Comput. Civil. Eng., Vol. 13, No. 2, pp. 103-109.
Alshibli, K. A. and Al-Hamdan, M. Z., 2001, “Estimating Volume Change of Triaxial Soil
Specimens from Planar Images,” Comput-Aided. Civ. Inf., Vol. 16, No. 6, pp. 415–421.
ASTM D698-12e2, 2012, "Standard Test Methods for Laboratory Compaction
Characteristics of Soil Using Standard Effort (12 400 ft-lbf/ft3 (600 kN-m/m3))",
ASTM International, West Conshohocken, PA, www.astm.org.
doi:10.1520/D0698-12E02.
ASTM D4767-11, 2011, "Standard Test Methods for Consolidated Undrained
Triaxial Compression Test for Cohesive Soils," ASTM International, West
Conshohocken, PA, www.astm.org. doi:10.1520/D4767-11.
Bésuelle, P. and Desrues, J., 2001, “An Internal Instrumentation for Axial and Radial Strain
Measurements in Triaxial Tests,” Geotech. Test J., Vol. 24, No. 2, pp. 193-199.
Bhandari, A. R., Powrie, W., and Harkness, R. M., 2012, “A Digital Image-Based Deformation
Measurement System for Triaxial Tests,” Geotech. Test. J., Vol. 35, No. 2, pp. 1-18.
Bishop, A. W. and Donald, I. B., 1961, “The Experimental Study of Partly Saturated Soil in
Triaxial Apparatus,” presented at the Fifth International Conference on Soil Mechanics
and Found Engineering, Paris, France, July 17-22, 1961, Dunod, Paris, pp. 13-21.
Cignoni, P., Rocchini, C., and Scopigno, R., 1998, “Metro: Measuring Error on Simplified
Surfaces,” Comp. Graph. Forum, Vol. 17, No. 2, pp. 167-174.
Clayton, C. R. I., Khatrush, S. A., Adriano, V. D. B., and Siddique, A., 1989, “The Use of Hall
Effect Semiconductors in Geotechnical Instrumentation,” Geotech. Test. J., Vol. 12,
No. 1, pp. 69-76.
Desrues, J., Chambon, R., Mokni, M., and Mazerolle, F., 1996, “Void Ratio Evolution Inside
Shear Bands in Triaxial Sand Specimens Studied by Computed Tomography.
Géotechnique, Vol. 46, No. 35, pp. 529-546.
Gachet, P., Geiser, F., Laloui, L., and Vulliet, L., 2007, “Automated Digital Image Processing
for Volume Change Measurement in Triaxial Cells,” Geotech. Test. J., Vol. 30, No. 2,
pp. 98-103.
Geomagic Design X; Version 17, 2015, “3D Computer-Aided Design (CAD) software,” 3D
Systems, Rock Hill, SC.
![Page 149: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/149.jpg)
133
Hormdee, D., Kaikeerati, N., and Jirawattana, P., 2014, “Application of Image Processing for
Volume Measurement in Multistage Triaxial Tests,” Adv. Mat. Res., Vol. 931-932,
pp. 501-505.
Kazhdan, M. and Hoppe, H., 2013, “Screened Poisson Surface Reconstruction,” ACM Trans.
Graph., Vol. 32, No. 3, pp. 1-13.
Kikkawa, N., Nakata, Y., Hyodo, M., Murata, H., and Nishio, S., 2006, "Three-Dimensional
Measurement of Local Strain Using Digital Stereo Photogrammetry in the Triaxial Test,"
Geomechanics and Geotechnics of Particulate Media, M. Hyodo, H. Murata, and Y.
Nakata, Eds., Taylor and Francis, London, pp. 61-67.
Leong, E.C., Agus, S.S., and Rahardjo, H., 2004, “Volume Change Measurement of Soil
Specimen in Triaxial Test,” Geotech. Test. J., Vol. 27, No. 1, pp. 1-10.
Li, L., Zhang, X., Chen, G., and Lytton, R., 2016, “Measuring Unsaturated Soil Deformations
During Triaxial Testing Using a Photogrammetry-Based Method,” Can. Geotech. J.,
Vol. 53, No. 3, pp. 472-489.
Macari, E. J., Parker, J. K., and Costes, N. C., 1997, “Measurement of Volume Changes in
Triaxial Tests Using Digital Imaging Techniques,” Geotech. Test. J., Vol. 20, No. 1,
pp. 103–109.
Mehdizadeh, A., Disfani, M. M., Evans, R., Arulrajah, A., and Ong, D. E. L., 2016, “Discussion
of ‘Development of an Internal Camera-Based Volume Determination System for
Triaxial Testing’ by S. E. Salazar, A. Barnes and R. A. Coffman
(doi:10.1520/GTJ20140249).” Geotech. Test J., Vol. 39, No. 1, pp. 165-168.
doi:10.1520/GTJ20150153.
Messerklinger, S. and Springman, S. M., 2007, “Local Radial Displacement Measurements of
Soil Specimens in a Triaxial Test Apparatus Using Laser Transducers,” Geotech. Test. J.,
Vol. 30, No. 6, pp. 1-12.
Mirtich, B., 1996, “Fast and Accurate Computation of Polyhedral Mass Properties,” J. Graph.
Tools, Vol. 1, No. 2, pp. 31-50.
Ng, C. W. W., Zhan L. T., and Cui Y. J., 2002, “A New Simple System for Measuring Volume
Changes in Unsaturated Soils,” Can. Geotech. J., Vol. 39, No. 3, pp. 757-764.
PhotoModeler Scanner; Version 2015.0.0, 2015, “3D Measurements and Models Software,” Eos
Systems, Vancouver, BC, Canada.
Rechenmacher, A. L. and Medina-Cetina, Z., 2007, “Calibration of Soil Constitutive Models
with Spatially Varying Parameters,” J. Geotech. Geoenviron., Vol. 133, No. 12,
pp. 1567–1576.
![Page 150: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/150.jpg)
134
Romero, E., Facio, J. A., Lloret, A., Gens, A., and Alonso, E. E., 1997, “A New Suction and
Temperature Controlled Triaxial Apparatus,” presented at the Fourteenth International
Conference on Soil Mechanics and Foundation Engineering, Hamburg, Germany,
September 6-12, 1997, August Aimé Balkema, Amsterdam, pp. 185-188.
Sachan, A. and Penumadu, D., 2007, “Strain Localization in Solid Cylindrical Clay Specimens
Using Digital Image Analysis (DIA) Technique,” Soils Found., Vol. 47, No. 1, pp. 67-78.
Salazar, S. E., Barnes, A., and Coffman, R. A., 2015, “Development of an Internal Camera
Based Volume Determination System for Triaxial Testing,” Geotech. Test. J., Vol. 38,
No. 4, pp. 549-555. doi:10.1520/GTJ20140249.
Salazar, S. E., Barnes, A., and Coffman, R. A., 2017, “Closure to “Discussion of ‘Development
of an Internal Camera-Based Volume Determination System for Triaxial Testing’ S. E.
Salazar, A. Barnes and R. A. Coffman” by A. Mehdizadeh, M. M. Disfani, R. Evans, A.
Arulrajah and D. E. L. Ong,” Geotech. Test. J., Vol. 40, No. 1, pp. 47-51.
doi:10.1520/GTJ20160154.
Salazar, S. E. and Coffman, R. A., 2015a, “Consideration of Internal Board Camera Optics for
Triaxial Testing Applications,” Geotech. Test. J., Vol. 38, No. 1, pp. 40-49.
doi:10.1520/GTJ20140163.
Salazar, S. E. and Coffman, R. A., 2015b, “Discussion of “A Photogrammetry-Based Method to
Measure Total and Local Volume Changes of Unsaturated Soils During Triaxial Testing”
by Zhang et al. (doi:10.1007/s11440-014-0346-8)” Acta Geotechnica, Vol. 10, No. 5,
pp. 693-696. doi:10.1007/s11440-015-0380-1.
Salazar, S. E. and Coffman, R. A., 2016, “Pressurized Fluid-Submerged, Internal, Close-Range
Photogrammetry System for Laboratory Testing,” U.S. Patent and Trademark Office.
Patent 15/360,820. Filed November, 2016.
Scholey, G. K., Frost, J. D., Lo Presti, C. F., and Jamiolkowski, M., 1995, “A Review of
Instrumentation for Measuring Small Strains During Triaxial Testing of Soil Specimens,”
Geotech. Test. J., Vol. 18, No. 2, pp. 137–156.
Uchaipichat, A., Khalili, N., and Zargarbashi, S., 2011, “A Temperature Controlled Triaxial
Apparatus for Testing Unsaturated Soils,” Geotech. Test. J., Vol. 34, No. 5, pp. 1-9.
Viggiani, G., Lenoir, N., Bésuelle, P., Di Michiel, M., Marello, S., Desrues, J., and Kretzschmer,
M., 2004, “X-ray Microtomography for Studying Localized Deformation in Fine-Grained
Geomaterials Under Triaxial Compression,” Cr Acad Sci II B-Mec, Vol. 332,
pp. 819-826.
Zhang, X., Li, L., Chen, G., and Lytton, R., 2015a, “A Photogrammetry-Based Method to
Measure Total and Local Volume Changes of Unsaturated Soils During Triaxial
Testing,” Acta Geotechnica, Vol. 10, No. 1, pp. 55-82. doi:10.1007/s11440-014-0346-8.
![Page 151: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/151.jpg)
135
Zhang, X., Li, L., Chen, G., and Lytton, R., 2015b, “Reply to “Discussion of ‘A
Photogrammetry-Based Method to Measure Total and Local Volume Changes of
Unsaturated Soils During Triaxial Testing’ by Zhang et al. (doi 10.1007/s11440-014-
0346-8)” by Salazar and Coffman (doi 10.1007/s11440-015-0380-1),” Acta Geotechnica,
Vol. 10, No. 5, pp. 697-702. doi: 10.1007/s11440-015-0384-x.
![Page 152: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/152.jpg)
136
CHAPTER 8: CONCLUSIONS
8.1. Chapter Overview
Conclusions that were drawn from this work are contained in this chapter. Highlights
from the work are summarized in Section 8.2. The limitations of the work are outlined in Section
8.3. Finally, recommendations for future work are put forth in Section 8.4.
8.2. Highlights
The internal-cell photogrammetry technique that was developed in this work was
successfully applied to the monitoring of soil specimens during triaxial compression and triaxial
extension tests. Three-dimensional models with watertight meshes were derived from
photogrammetric processing. Measurements of total volume were obtained from each model.
The models developed for the specimen, when the specimen had been subjected to different
levels of axial stress, were differenced to quantify the development of axial, radial, and localized
deformations on the specimen surface during testing. Each set of measurements that were used to
derive a specimen volume was collected and processed independently. Therefore, the obtained
volume did not rely on assumptions of specimen dimensions or the deformation behavior during
the testing. This measurement technique reduced the source of systematic error in the
determination of specimen volume.
Volume measurements were made for a rigid reference (dummy) specimen, using five
independent techniques, to evaluate the internal-cell photogrammetry technique. Results for the
internal-cell photogrammetry technique fell within 0.13 percent of the reference technique and
within 0.50 percent of all other techniques. A sensitivity study was performed to examine the
effect that the number of photographs used in processing had on the total volume of the
specimen. The study revealed that a reduced number of photographs (from 320 to 80
![Page 153: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/153.jpg)
137
photographs) was suitable for accurately reconstructing a specimen. Accordingly, this reduction
in photographs resulted in reduced processing time.
8.3. Limitations
Chapter 7 contained a detailed discussion of the limitations of the internal cell
photogrammetry technique and the sources of error that were identified. For reference, these
limitations included:
the precision of repeat interval stops (Section 7.6.8.1);
the refinement of specimen models (Section 7.6.8.2);
the accuracy of external geometry measurements (Section 7.6.8.3); and
the determination of specimen ends (Section 7.6.8.4).
In addition to the limitations discussed in Chapter 7, the following limitations of the work were
identified:
the lack of drained triaxial tests performed as part of this work;
the lack of repeat tests performed as part of this work;
the inability of the photogrammetry software to automatically identify all targets;
the amount of time required for photogrammetric processing of results;
the low resolution of the board camera devices; and
the poor performance of the board camera devices in low light conditions.
8.4. Recommendations
In accordance with the limitations that were identified in the previous section, the
following recommendations are made for future work.
The precision of the camera towers during repeat occupation should be improved with a
mechanized, rotating track (e.g. with a small stepper motor). This mechanization could
![Page 154: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/154.jpg)
138
also foreseeably decrease the amount of time that the triaxial test is paused for during
acquisition of photographs.
The number/density of targets on the surface of the specimen should be increased. The
increase in density will increase the number of surveyed points and therefore also
increase the resolution of the three-dimensional model.
Although a comparison with other photogrammetric methods in the literature indicated
that the technique presented in this work was no more computationally costly, a reduction
in processing time is still desirable. It is therefore recommended that an alternative
(preferably) open-source software package be implemented for photogrammetric
processing that would allow for improved automatic target and texture detection.
Incorporating an improved board camera design in future iterations of the system could
foreseeably improve image resolution and sensor performance in low light conditions,
thereby improving the quality of the image data.
The strain measurements collected with this technique should be used in conjunction with
finite element analysis to predict the stress distribution throughout the soil specimen.
Such an inverse solution will help to develop a better understanding of the constitutive
behavior of the soil.
![Page 155: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/155.jpg)
139
CHAPTER 9: WORKS CITED
3D Systems, Inc., 2015, Geomagic Design X (Version 17). 3D Computer-Aided Design
(CAD) Software. http://www.geomagic.com/en/products/design/overview
Adrian, R. J., 1991, "Particle Imaging Techniques for Experimental Fluid Mechanics," Ann. Rev.
Fluid Mech., Vol. 23, No. 1, pp. 261-304.
Airy, G. B., 1835, “On the Diffraction of an Object-Glass With Circular Aperture,” Trans.
Cambr. Philos. Soc., Vol. 5, pp. 283–291.
Alshibli, K. A. and Sture, S., 1999, “Sand Shear Band Thickness Measurements by Digital
Imaging Techniques,” J. Comput. Civil. Eng., Vol. 13, No. 2, pp. 103-109.
Alshibli, K. A. and Al-Hamdan, M. Z., 2001, “Estimating Volume Change of Triaxial Soil
Specimens from Planar Images,” Comput-Aided. Civ. Inf., Vol. 16, No. 6, pp. 415–421.
ASTM D698-12e2, 2012, "Standard Test Methods for Laboratory Compaction
Characteristics of Soil Using Standard Effort (12 400 ft-lbf/ft3 (600 kN-m/m3))",
ASTM International, West Conshohocken, PA, www.astm.org.
doi:10.1520/D0698-12E02.
ASTM D4767-11, 2011, "Standard Test Methods for Consolidated Undrained
Triaxial Compression Test for Cohesive Soils," ASTM International, West
Conshohocken, PA, www.astm.org. doi:10.1520/D4767-11.
Aversa, S. and Nicotera, M. V., 2002, "A Triaxial and Oedometer Apparatus for Testing
Unsaturated Soils," Geotech. Test. J., Vol. 25, No. 1, pp. 3-15.
Bésuelle, P. and Desrues, J., 2001, “An Internal Instrumentation for Axial and Radial Strain
Measurements in Triaxial Tests,” Geotech. Test J., Vol. 24, No. 2, pp. 193-199.
Bhandari, A. R., Powrie, W., and Harkness, R. M., 2012, “A Digital Image-Based Deformation
Measurement System for Triaxial Tests,” Geotech. Test. J., Vol. 35, No. 2, pp. 1-18.
Bishop, A. W. and Donald, I. B., 1961, “The Experimental Study of Partly Saturated Soil in
Triaxial Apparatus,” presented at the Fifth International Conference on Soil Mechanics
and Found Engineering, Paris, France, July 17-22, 1961, Dunod, Paris, pp. 13-21.
Cignoni, P., Rocchini, C., and Scopigno, R., 1998, “Metro: Measuring Error on Simplified
Surfaces,” Comp. Graph. Forum, Vol. 17, No. 2, pp. 167-174.
Clayton, C. R. I. and Khatrush, S. A., 1986, "A New Device for Measuring Local Axial Strains
on Triaxial Specimens," Géotechnique, Vol. 36, No. 4, pp. 593-597.
![Page 156: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/156.jpg)
140
Clayton, C. R. I., Khatrush, S. A., Adriano, V. D. B., and Siddique, A., 1989, “The Use of Hall
Effect Semiconductors in Geotechnical Instrumentation,” Geotech. Test. J., Vol. 12,
No. 1, pp. 69-76.
Cuccovillo, T. and Coop, M. R., 1997, “The Measurement of Local Axial Strains in Triaxial
Tests Using LVDTs,” Géotechnique, Vol. 47, No. 1, pp. 167–171.
Cui, Y.J. and Delage, P., 1996, "Yielding and Plastic Behavior of an Unsaturated Compacted
Silt," Géotechnique, Vol. 46, No. 2, pp. 291-311.
Desrues, J., Chambon, R., Mokni, M., and Mazerolle, F., 1996, “Void Ratio Evolution Inside
Shear Bands in Triaxial Sand Specimens Studied by Computed Tomography.
Géotechnique, Vol. 46, No. 35, pp. 529-546.
Ehrgott, J. Q., 1971, "Calculation of Stress and Strain from Triaxial Test Data no Undrained Soil
Specimens," Miscellaneous Paper S-71-9, U.S. Army Engineer Waterways Experiment
Station, Vicksburg, Mississippi.
Eos Systems Inc., 2014, PhotoModeler (Version 2014.1.1). 3D Measurements and Models
Software. http://www.photomodeler.com/index.html (Last accessed 13 Oct 2014).
Eos Systems, Inc., 2015, PhotoModeler Scanner (Version 2015.0.0). 3D Measurements and
Models Software. http://www.photomodeler.com/products/scanner/default.html
Fresnel, A.-J., 1819, “Memoir on the Diffraction of Light,” The Wave Theory of Light, H. Crew,
Trans., American Book Company, New York.
Furukawa, Y. and Hernández, C., 2013, “Multi-View Stereo: A Tutorial,” Found. Trends
Comput. Graphics Vis., Vol. 9, Nos. 1–2, pp. 1–148.
Gachet, P., Geiser, F., Laloui, L., and Vulliet, L., 2007, “Automated Digital Image Processing
for Volume Change Measurement in Triaxial Cells,” Geotech. Test. J., Vol. 30, No. 2,
pp. 98-103.
Geomagic Design X; Version 17, 2015, “3D Computer-Aided Design (CAD) software,” 3D
Systems, Rock Hill, SC.
Grimaldi, F. M., 1665, Physico–mathesis de lumine, coloribus, et iride, aliisque adnexis libri duo
[A physiomathematical thesis on light, colors, the rainbow and other related topics in two
books], Bologna, Italy (in Latin).
Gudehus, G. and Nübel, K., 2004, “Evolution of Shear Bands in Sand,” Géotechnique, Vol. 54,
No. 3, pp. 187–201.
Guler, M., Edil, T. B., and Bosscher, P. J., 1999, Measurement of Particle Movement in Granular
Soils Using Image Analysis," J. Comput. Civ., Vol. 13, No. 2, pp. 116-122.
![Page 157: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/157.jpg)
141
Hendriks, B. H. W., Kuiper, S., van As, M. A. J., Renders, C. A., and Tukker, T. W., 2006,
“Variable Liquid Lenses for Electronic Products,” Proc. SPIE, Vol. 6034, pp. 10–18.
Herschel, J. F. W., 1835, Treatise on Astronomy, 3rd ed., Carey, Lea & Blanchard,
Philadelphia, PA.
Hormdee, D., Kaikeerati, N., and Jirawattana, P., 2014, “Application of Image Processing for
Volume Measurement in Multistage Triaxial Tests,” Adv. Mat. Res., Vol. 931-932,
pp. 501-505.
Iskander, M. and Liu, J., 2010, "Spatial Deformation Measurement Using Transparent Soil,"
Geotech. Test. J., Vol. 33, No. 4, 8 pgs.
Jain, S., Wang, Y. H., and Fredlund, D. G., 2015, "Non-Contact Sensing System to Measure
Specimen Volume During Shrinkage Test," Geotech. Test. J., Vol. 38, No. 6,
pp. 936-949. doi:10.1520/GTJ20140274.
Kazhdan, M. and Hoppe, H., 2013, “Screened Poisson Surface Reconstruction,” ACM Trans.
Graph., Vol. 32, No. 3, pp. 1-13.
Khan, M. H. and Hoag, D. L., 1979, “A Noncontacting Transducer for Measurement of Lateral
Strains,” Can. Geotech. J., Vol. 16, No. 2, pp. 409–411.
Kikkawa, N., Nakata, Y., Hyodo, M., Murata, H., and Nishio, S., 2006, "Three-Dimensional
Measurement of Local Strain Using Digital Stereo Photogrammetry in the Triaxial Test,"
Geomechanics and Geotechnics of Particulate Media, M. Hyodo, H. Murata, and Y.
Nakata, Eds., Taylor and Francis, London, pp. 61-67.
Kingslake, R. and Johnson, R. B., 2010, Lens Design Fundamentals, 2nd ed., Academic Press,
New York.
Kuiper, S. and Hendriks, B. H. W., 2004, “Variable-Focus Liquid Lens for Miniature Cameras,”
Appl. Phys. Lett., Vol. 85, No. 7, pp. 1128–1130.
Kuo, C.-H. and Ye, Z., 2004, “Sonic Crystal Lenses that Obey the Lensmaker’s Formula,”
J. Phys. D., Vol. 37, No. 15, pp. 2155–2159.
Lasance, C. and Simons, R., 2005, “Advances in High Performance Cooling for Electronics,”
Electron. Cool., Vol. 11, No. 4, pp. 22–39.
Laudo, J. S., Wurm, K., and Dodson, C., 1998, “Liquid-Filled Underwater Camera Lens
System,” Proc. SPIE, Vol. 3482, pp. 698–702.
Leong, E.C., Agus, S.S., and Rahardjo, H., 2004, “Volume Change Measurement of Soil
Specimen in Triaxial Test,” Geotech. Test. J., Vol. 27, No. 1, pp. 47-56.
![Page 158: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/158.jpg)
142
Li, L., Zhang, X., Chen, G., and Lytton, R., 2016, “Measuring Unsaturated Soil Deformations
During Triaxial Testing Using a Photogrammetry-Based Method,” Can. Geotech. J.,
Vol. 53, No. 3, pp. 472-489.
Liang, L., Saada, A., Figueroa, J. L., and Cope, C. T., 1997, “The Use of Digital Image
Processing in Monitoring Shear Band Development,” Geotech. Test. J., Vol. 20, No. 3,
pp. 324–339.
Macari, E. J., Parker, J. K., and Costes, N. C., 1997, “Measurement of Volume Changes in
Triaxial Tests Using Digital Imaging Techniques,” Geotech. Test. J., Vol. 20, No. 1,
pp. 103–109.
Mahajan, V. N., 1998, Optical Imaging and Aberrations: Part I. Ray Geometrical Optics, SPIE,
Bellingham, WA.
Miramontes, L. D., 2016, "Validation of an Internal Camera Based Volume Determination
System for Triaxial Testing," Civil Engineering Undergraduate Honors Theses. 33.
Mohapatra, S. C. and Loikits, D., 2005, “Advances in Liquid Coolant Technologies for
Electronics Cooling,” Proceedings of the 21st IEEE Semiconductor Thermal
Measurement and Management Symposium, San Jose, CA, March 15–17, pp. 354–360.
Mehdizadeh, A., Disfani, M. M., Evans, R., Arulrajah, A., and Ong, D. E. L., 2016, “Discussion
of ‘Development of an Internal Camera-Based Volume Determination System for
Triaxial Testing’ by S. E. Salazar, A. Barnes and R. A. Coffman
(doi:10.1520/GTJ20140249).” Geotech. Test J., Vol. 39, No. 1, pp. 165-168.
doi:10.1520/GTJ20150153.
Messerklinger, S. and Springman, S. M., 2007, “Local Radial Displacement Measurements of
Soil Specimens in a Triaxial Test Apparatus Using Laser Transducers,” Geotech. Test. J.,
Vol. 30, No. 6, pp. 1-12.
Mirtich, B., 1996, “Fast and Accurate Computation of Polyhedral Mass Properties,” J. Graph.
Tools, Vol. 1, No. 2, pp. 31-50.
Newton, I., 1730, Opticks: Or a Treatise of the Reflections, Refractions, Inflections and Colours
of Light, 4th ed., William Innys, London.
Nguyen, N.-T., 2010, “Micro-Optofluidic Lenses: A Review,” Biomicrofluidics, Vol. 4, No. 3,
031501.
Ng, C. W. W. and Chiu, C. F., 2001, “Behaviour of a Loosely Compacted Unsaturated Volcanic
Soil,” J. Geotech. Geoenviron. Eng., Vol. 127, No. 12, pp. 1027–1036.
Ng, C. W. W., Zhan L. T., and Cui Y. J., 2002, “A New Simple System for Measuring Volume
Changes in Unsaturated Soils,” Can. Geotech. J., Vol. 39, No. 3, pp. 757-764.
![Page 159: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/159.jpg)
143
Nübel, K. and Weitbrecht, V., 2002, “Visualization of Localization in Grain Skeletons With
Particle Image Velocimetry,” J. Testing Eval., Vol. 30, No. 4, pp. 322–329.
Önal, O., Ören, A. H., Özden, G., and Kaya, A., 2008, “Determination of Cylindrical Soil
Specimen Dimensions by Imaging With Applications to Volume Change of Bentonite–
Sand Mixtures,” Geotech. Test. J., Vol. 31, No. 2, pp. 124–131.
Ören, A. H., Önal, O., Özden, G., and Kaya, A., 2006, “Nondestructive Evaluation of
Volumetric Shrinkage of Compacted Mixtures Using Digital Image Analysis,”
Eng. Geol., Vol. 85, No. 3, pp. 239–250.
Parker, J. K., 1987, “Image Processing and Analysis for the Mechanics of Granular Materials
Experiment,” ASME Proceedings of the 19th SE Symposium on System Theory, ASME,
New York, pp. 536–541.
Pinyol, N. M. and Alvarado, M., 2017, "Novel Analysis for Large Strains Based on Particle
Image Velocimetry," Can. Geotech. J., Vol. 54, No. 7, pp. 933-944. doi:10.1139/cgj-
2016-0327.
Renner, E., 2000, Pinhole Photography: Rediscovering a Historic Technique, 2nd ed.,
Butterworth–Heinemann, Boston, MA.
Roichman, Y., Waldron, A., Gardel, E., and Grier, D. G., 2006, “Optical Traps with Geometric
Aberrations,” Appl. Opt., Vol. 45, No. 15, pp. 3425–3429.
PhotoModeler Scanner; Version 2015.0.0, 2015, “3D Measurements and Models Software,” Eos
Systems, Vancouver, BC, Canada.
Puppala, A. J., Katha, B., and Hoyos, L. R., 2004, “Volumetric Shrinkage Strain Measurements
in Expansive Soils Using Digital Imaging Technology,” Geotech. Test. J., Vol. 27,
No. 6, pp. 547–556.
Race, M. L. and Coffman, R. A., 2011, “Effects of Piston Uplift, Piston Friction, and
Machine Deflection in Reduced Triaxial Extension Testing,” ASCE Geotechnical
Special Publication No. 211, Proceedings GeoFrontiers 2011: Advances in
Geotechnical Engineering, ASCE, Reston, VA, pp. 2649–2658.
Rechenmacher, A. L. and Finno, R. J., 2004, “Digital Image Correlation to Evaluate Shear
Banding in Dilative Sands,” Geotech. Test. J., Vol. 27, No. 1, pp. 13–22.
Rechenmacher, A. L. and Medina-Cetina, Z., 2007, “Calibration of Soil Constitutive Models
with Spatially Varying Parameters,” J. Geotech. Geoenviron., Vol. 133, No. 12,
pp. 1567–1576.
![Page 160: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/160.jpg)
144
Romero, E., Facio, J. A., Lloret, A., Gens, A., and Alonso, E. E., 1997, “A New Suction and
Temperature Controlled Triaxial Apparatus,” presented at the Fourteenth International
Conference on Soil Mechanics and Foundation Engineering, Hamburg, Germany,
September 6-12, 1997, August Aimé Balkema, Amsterdam, pp. 185-188.
Sachan, A. and Penumadu, D., 2007, “Strain Localization in Solid Cylindrical Clay Specimens
Using Digital Image Analysis (DIA) Technique,” Soils Found., Vol. 47, No. 1, pp. 67-78.
Salazar, S. E., Barnes, A., and Coffman, R. A., 2015, “Development of an Internal Camera
Based Volume Determination System for Triaxial Testing,” Geotech. Test. J., Vol. 38,
No. 4, pp. 549-555. doi:10.1520/GTJ20140249.
Salazar, S. E., Barnes, A., and Coffman, R. A., 2017, “Closure to “Discussion of ‘Development
of an Internal Camera-Based Volume Determination System for Triaxial Testing’ S. E.
Salazar, A. Barnes and R. A. Coffman” by A. Mehdizadeh, M. M. Disfani, R. Evans, A.
Arulrajah and D. E. L. Ong,” Geotech. Test. J., Vol. 40, No. 1, pp. 47-51.
doi:10.1520/GTJ20160154.
Salazar, S. E. and Coffman, R. A., 2014, “Design and Fabrication of End Platens for
Acquisition of Small-Strain Piezoelectric Measurements During Large-Strain
Triaxial Extension and Triaxial Compression Testing,” Geotech. Test. J., Vol. 37,
No. 6, pp. 948–958. doi:10.1520/GTJ20140057.
Salazar, S. E. and Coffman, R. A., 2015, “Consideration of Internal Board Camera Optics for
Triaxial Testing Applications,” Geotech. Test. J., Vol. 38, No. 1, pp. 40-49.
doi:10.1520/GTJ20140163.
Salazar, S. E. and Coffman, R. A., 2015, “Discussion of “A Photogrammetry-Based Method to
Measure Total and Local Volume Changes of Unsaturated Soils During Triaxial Testing”
by Zhang et al. (doi:10.1007/s11440-014-0346-8)” Acta Geotech., Vol. 10, No. 5,
pp. 693-696. doi:10.1007/s11440-015-0380-1.
Salazar, S. E. and Coffman, R. A., 2016, “Pressurized Fluid-Submerged, Internal, Close-Range
Photogrammetry System for Laboratory Testing,” U.S. Patent and Trademark Office.
Patent 15/360,820. Filed November, 2016.
Salazar, S. E., Miramontes, L. D., Barnes, A., Bernhardt, M. L., and Coffman, R. A., 2017,
“Verification of an Internal Close-Range Photogrammetry Approach for Volume
Determination During Triaxial Testing,” Geotech. Test. J. Submitted for Review.
Manuscript Number: GTJ-2017-0125.R2.
Scholey, G. K., Frost, J. D., Lo Presti, C. F., and Jamiolkowski, M., 1995, “A Review of
Instrumentation for Measuring Small Strains During Triaxial Testing of Soil Specimens,”
Geotech. Test. J., Vol. 18, No. 2, pp. 137–156.
Schmidt, R., 2005, “Liquid Cooling is Back,” Electron. Cool., Vol. 11, No. 3, pp. 34–38.
![Page 161: Development of an Internal Camera-Based Volume](https://reader035.vdocument.in/reader035/viewer/2022081511/62a492199e066567323928e5/html5/thumbnails/161.jpg)
145
Stanier, S. A., Blaber, J., Take, W. A., and White, D. J., 2016, "Improved Image-Based
Deformation Measurement for Geotechnical Applications," Can Geotech. J., Vol. 53,
No. 5, pp. 727-739. doi:10.1139/cgj-2015-0253.
Strutt, J. W., Lord Rayleigh, 1891, “On Pin-Hole Photography,” Philos. Mag., Vol. 31,
pp. 87–99.
Toyota, H., Sakai, N., and Nishimura, T., 2001, "Effects of Stress History due to Unsaturation
and Drainage Conditions on Shear Properties of Unsaturated Cohesive Soil," Soils
Found., Vol. 41, No. 1, pp. 13-24.
Uchaipichat, A., Khalili, N., and Zargarbashi, S., 2011, “A Temperature Controlled Triaxial
Apparatus for Testing Unsaturated Soils,” Geotech. Test. J., Vol. 34, No. 5, pp. 1-9.
Viggiani, G., Lenoir, N., Bésuelle, P., Di Michiel, M., Marello, S., Desrues, J., and Kretzschmer,
M., 2004, “X-ray Microtomography for Studying Localized Deformation in Fine-Grained
Geomaterials Under Triaxial Compression,” Cr Acad Sci II B-Mec, Vol. 332,
pp. 819-826.
Wheeler, S. J., 1986, "The Stress-Strain Behaviour of Soils Containing Gas Bubbles," Ph.D.
dissertation, Oxford University, Oxford, UK.
White, D., Take, W., and Bolton, M., 2003, "Soil Deformation Measurement Using Particle
Image Velocimetry (PIV) and Photogrammetry," Géotechnique, Vol. 53, No. 7,
pp. 619-631.
Young, T., 1802, “Bakerian Lecture: On the Theory of Light and Colours,” Philos. Trans. R.
Soc. Lond., Vol. 92, pp. 12–48.
Young, M., 1971, “Pinhole Optics,” Appl. Opt., Vol. 10, No. 12, pp. 2763–2767.
Zhang, X., Li, L., Chen, G., and Lytton, R., 2015, “A Photogrammetry-Based Method to
Measure Total and Local Volume Changes of Unsaturated Soils During Triaxial
Testing,” Acta Geotechnica, Vol. 10, No. 1, pp. 55-82. doi:10.1007/s11440-014-0346-8.
Zhang, X., Li, L., Chen, G., and Lytton, R., 2015, “Reply to “Discussion of ‘A
Photogrammetry-Based Method to Measure Total and Local Volume Changes of
Unsaturated Soils During Triaxial Testing’ by Zhang et al. (doi 10.1007/s11440-014-
0346-8)” by Salazar and Coffman (doi 10.1007/s11440-015-0380-1),” Acta Geotechnica,
Vol. 10, No. 5, pp. 697-702. doi: 10.1007/s11440-015-0384-x.
Zhao, Y. and Coffman, R. A., 2016, “Back-Pressure Saturated Constant-Rate-of-Strain
Consolidation Device With Bender Elements: Verification of System Compliance,”
J. Test. Eval., Vol. 44, No. 6, pp. 2375–2386.
Zhao, Y., Mahmood, N. S., and Coffman, R. A., 2017, “Small-Strain and Large-Strain
Modulus Measurements with a Consolidation Device,” J. Test. Eval. (In Press).